Red Hat Enterprise Linux 7 Beta Performance Tuning Guide en US

Red Hat Enterprise Linux 7.
2 Beta
Performance Tuning Guide
Optimizing subsystem throughput in Red Hat Enterprise Linux 7
Laura Bailey
Red Hat Subject Matter Experts
Red Hat Enterprise Linux 7.2 Beta Performance Tuning Guide
Optimizing subsystem throughput in Red Hat Enterprise Linux 7

Laura Bailey
[email protected] m
Red Hat Subject Matter Experts
Legal Notice
Co pyright 20 14 20 15 Red Hat, Inc. and o thers.
This do cument is licensed by Red Hat under the Creative Co mmo ns Attributio n-ShareAlike 3.0
Unpo rted License. If yo u distribute this do cument, o r a mo dified versio n o f it, yo u must pro vide
attributio n to Red Hat, Inc. and pro vide a link to the o riginal. If the do cument is mo dified, all Red
Hat trademarks must be remo ved.
Red Hat, as the licenso r o f this do cument, waives the right to enfo rce, and agrees no t to assert,
Sectio n 4 d o f CC-BY-SA to the fullest extent permitted by applicable law.
Red Hat, Red Hat Enterprise Linux, the Shado wman lo go , JBo ss, MetaMatrix, Fedo ra, the Infinity
Lo go , and RHCE are trademarks o f Red Hat, Inc., registered in the United States and o ther
co untries.
Linux is the registered trademark o f Linus To rvalds in the United States and o ther co untries.
Java is a registered trademark o f Oracle and/o r its affiliates.
XFS is a trademark o f Silico n Graphics Internatio nal Co rp. o r its subsidiaries in the United
States and/o r o ther co untries.
MySQL is a registered trademark o f MySQL AB in the United States, the Euro pean Unio n and
o ther co untries.
No de.js is an o fficial trademark o f Jo yent. Red Hat So ftware Co llectio ns is no t fo rmally
related to o r endo rsed by the o fficial Jo yent No de.js o pen so urce o r co mmercial pro ject.
The OpenStack Wo rd Mark and OpenStack Lo go are either registered trademarks/service
marks o r trademarks/service marks o f the OpenStack Fo undatio n, in the United States and o ther
co untries and are used with the OpenStack Fo undatio n's permissio n. We are no t affiliated with,
endo rsed o r spo nso red by the OpenStack Fo undatio n, o r the OpenStack co mmunity.
All o ther trademarks are the pro perty o f their respective o wners.
Abstract
The Red Hat Enterprise Linux 7 Perfo rmance Tuning Guide explains ho w to o ptimize Red Hat
Enterprise Linux 7 perfo rmance. It also do cuments perfo rmance-related upgrades in Red Hat
Enterprise Linux 7. The Perfo rmance Tuning Guide presents o nly field-tested and pro ven
pro cedures. No netheless, all pro spective co nfiguratio ns sho uld be set up and tested in a
testing enviro nment befo re being applied to a pro ductio n system. Backing up all data and
co nfiguratio n settings prio r to tuning is also reco mmended. No te: this do cument is under active
develo pment; is subject to substantial change; and is pro vided o nly as a preview. The included
info rmatio n and instructio ns sho uld no t be co nsidered co mplete and sho uld be used with
cautio n.
T able of Cont ent s
T able of Contents
. .hapt
C
. . . .er
. .1. .. Performance
. . . . . . . . . . . .Feat
. . . ures
. . . . .in
. .Red
. . . .Hat
. . . .Ent
. . .erprise
. . . . . . Linux
. . . . . .7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. . . . . . . . . .
1.1. New in 7.1
3
1.2. New in 7.0
3
. .hapt
C
. . . .er
. .2. .. Performance
. . . . . . . . . . . .Monit
. . . . .oring
. . . . .T. ools
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. . . . . . . . . .
2 .1. /p ro c
5
2 .2. G NO ME Sys tem Mo nito r
5
2 .3. Perfo rmanc e Co -Pilo t (PCP)
6
2 .4. Tuna
6
2 .5. Built in c o mmand line to o ls
6
2 .6 . tuned and tuned -ad m
7
2 .7. p erf
8
2 .8 . turb o s tat
8
2 .9 . io s tat
9
2 .10 . irq b alanc e
9
2 .11. s s
2 .12. numas tat
9
9
2 .13. numad
2 .14. Sys temTap
10
10
2 .15. O Pro file

2 .16 . Valg rind
10
11
. .hapt
C
. . . .er
. .3.
. .CPU
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1. 2. . . . . . . . . .
3 .1. Co ns id eratio ns
3 .2. Mo nito ring and d iag no s ing p erfo rmanc e p ro b lems
12
17
3 .3. Co nfig uratio n s ug g es tio ns
18
. .hapt
C
. . . .er
. .4. .. Memory
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2. 4. . . . . . . . . .
24
4 .3. Co nfig uratio n to o ls
25
28
. .hapt
C
. . . .er
. .5.
. .St
. .orage
. . . . . and
. . . . File
. . . .Syst
. . . .ems
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
...........
33
39
41
. .hapt
C
. . . .er
. .6. .. Net
. . . working
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52
...........
52
53
54
. .ppendix
A
. . . . . . . A.
..T
. .ool
. . . Reference
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6. 1. . . . . . . . . .
A .1. irq b alanc e
61
A .2. Tuna
62
A .3. ethto o l
64
A .4. s s
64
A .5. tuned
A .6 . tuned -ad m
A .7. p erf
A .8 . Perfo rmanc e Co -Pilo t (PCP)
64
65
66
67
A .9 . vms tat
A .10 . x8 6 _energ y_p erf_p o lic y
A .11. turb o s tat
67
68
69
Performance T uning G uide

A .11. turb o s tat
A .12. numas tat
A .13. numac tl
69
70
71
A .14. numad
A .15. O Pro file
A .16 . tas ks et
A .17. Sys temTap
72
74
75
75
. .ppendix
A
. . . . . . . B.
. . .Revision
. . . . . . . .Hist
. . . ory
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7. 6. . . . . . . . . .
Chapt er 1 . Performance Feat ures in Red Hat Ent erprise Linux 7
Chapter 1. Performance Features in Red Hat Enterprise Linux 7

Read this section for a brief overview of the performance-related changes included in Red Hat
Enterprise Linux 7.
1.1. New in 7.1

The kernel idle balance mechanism is now less likely to delay recently queued tasks on the same
CPU, reducing processing latency when idle balancing is in use.
The swapo n command has a new --discard parameter. This allows administrators to select
which discard policy to use when swapping, providing greater flexibility, and allowing improved
performance on Solid State D isks. By default, the --discard parameter discards the entire swap
area when swapo n is run, and discards free swap pages before reuse while swapping. For further
information, see the swapo n manual page: man swapo n.
tuned profiles can now be configured to alter kernel command line parameters by specifying the
cmdline parameter and a list of kernel parameters in the tuned profile. Any kernel parameters
specified with cmdline take effect after reboot.
1.2. New in 7.0

This guide has been completely rewritten and restructured for Red Hat Enterprise Linux 7.
d ead l i ne replaces cfq as the default I/O scheduler in Red Hat Enterprise Linux 7. This change
provides better performance for most common use cases.
The XFS file system replaces ext4 as the default file system, and is now supported to a maximum
file system size of 500 TB, and a maximum file offset of 8 EB (sparse files). Tuning
recommendations for XFS have been updated to assist with clarity.
The ext4 file system is now supported to a maximum file system size of 50 TB and a maximum file
size of 16 TB. Tuning recommendations have been updated accordingly. Additionally, support for
the ext2 and ext3 file systems is now provided by the ext4 driver.
The btrfs file system is now provided as a Technology Preview.
Red Hat Enterprise Linux 7 includes some minor performance improvements for GFS2.
T u n a has been updated to include support for configuration files, and adding and saving t u n ed
profiles. This updated version uses event-based sampling to consume fewer processor resources.
The graphical version has also been updated to allow realtime monitoring. T u n a is now
documented in Section 2.4, Tuna , Section 3.3.8, Configuring CPU, thread, and interrupt affinity
with Tuna , and Section A.2, Tuna .
The default profile for t u n ed is now thro ug hput-perfo rmance. This replaces the now removed
enterpri se-sto rag e profile. Several new profiles for networking and virtualization have been
added. Additionally, t u n ed now provides shell script callout and includes functionality.
The t u n ed - ad m tool now provides the reco mmend sub-command, which recommends an
appropriate tuning profile for your system. This also sets the default profile for your system at
install time, so can be used to return to the default profile.
Red Hat Enterprise Linux 7 provides support for automatic NUMA balancing. The kernel now
automatically detects which memory pages process threads are actively using, and groups the
threads and their memory into or across NUMA nodes. The kernel reschedules threads and
migrates memory to balance the system for optimal NUMA alignment and performance.
The performance penalty to enabling file system barriers is now negligible (less than 3% ). As
such, t u n ed profiles no longer disable file system barriers.
OProfile adds support for profiling based on the Linux Performance Events subsystem with the
new o perf tool. This new tool can be used to collect data in place of the o pco ntro l daemon.
Control groups remain available as a method of allocating resources to certain groups of
processes on your system. For detailed information about implementation in Red Hat
Enterprise Linux 7, see the Red Hat Enterprise Linux 7 Resource Management Guide, available from
http://access.redhat.com/site/documentation/Red_Hat_Enterprise_Linux/.
Chapt er 2 . Performance Monit oring T ools
Chapter 2. Performance Monitoring Tools

This chapter briefly describes some of the performance monitoring and configuration tools available
for Red Hat Enterprise Linux 7. Where possible, this chapter directs readers to further information
about how to use the tool, and examples of real life situations that the tool can be used to resolve.
The following knowledge base article provides a more comprehensive list of performance monitoring
tools suitable for use with Red Hat Enterprise Linux: https://access.redhat.com/site/solutions/173863.
2.1. /proc
The /pro c " file system" is a directory that contains a hierarchy of files that represent the current state
of the Linux kernel. It allows users and applications to see the kernel's view of the system.
The /pro c directory also contains information about system hardware and any currently running
processes. Most files in the /pro c file system are read-only, but some files (primarily those in
/proc/sys) can be manipulated by users and applications to communicate configuration changes to
the kernel.
For further information about viewing and editing files in the /pro c directory, refer to the Red Hat
Enterprise Linux 7 System Administrator's Reference Guide, available from
2.2. GNOME Syst em Monit or

The GNOME desktop environment includes a graphical tool, System Monitor, to assist you in
monitoring and modifying the behavior of your system. System Monitor displays basic system
information and allows you to monitor system processes and resource or file system usage.
System Monitor has four tabs, each of which displays different information about the system.
S yst em
This tab displays basic information about the system's hardware and software.
P ro cesses
This tab displays detailed information about active processes and the relationships
between those processes. The processes displayed can be filtered to make certain
processes easier to find. This tab also lets you perform some actions on the processes
displayed, such as start, stop, kill, and change priority.
R eso u rces
This tab displays the current CPU time usage, memory and swap space usage, and network
usage.
F ile Syst ems
This tab lists all mounted file systems, and provides some basic information about each,
such as the file system type, mount point, and memory usage.
To start System Monitor, press the Super key to enter the Activities Overview, type System Monitor, and
then press Enter.
For more information about System Monitor, see either the Help menu in the application, or the
Red Hat Enterprise Linux 7 System Administrator's Guide, available from
2.3. Performance Co-Pilot (PCP)

Red Hat Enterprise Linux 7 introduces support for Performance Co-Pilot (PCP), a suite of tools,
services, and libraries for acquiring, storing, and analysing system-level performance measurements.
Its light-weight, distributed architecture makes it particularly well suited to centralized analysis of
complex systems. Performance metrics can be added using the Python, Perl, C++ and C interfaces.
Analysis tools can use the client APIs (Python, C++, C) directly, and rich web applications can
explore all available performance data using a JSON interface.
The pcp package provides the command line tools and underlying functionality. The graphical tool
also requires the pcp-gui package.
The Red Hat Customer Portal has a number of useful articles to help you get started with PCP. An
index of these articles is available here: https://access.redhat.com/articles/1145953.
Additionally, the pcp-doc package provides comprehensive documentation, which is installed to
/usr/share/d o c/pcp-d o c by default. PCP also provides a man page for every tool; type man
toolname at the command line to view the man page for that tool.
2.4 . T una
Tuna adjusts configuration details such as scheduler policy, thread priority, and CPU and interrupt
affinity. The tuna package provides a command line tool and a graphical interface with equivalent
functionality.
Section 3.3.8, Configuring CPU, thread, and interrupt affinity with Tuna describes how to configure
your system with Tuna on the command line. For details about how to use Tuna, see Section A.2,
Tuna or the man page:
$ man tuna
2.5. Built in command line t ools

Red Hat Enterprise Linux 7 provides several tools that can be used to monitor your system from the
command line, allowing you to monitor your system outside run level 5. This chapter discusses each
tool briefly and provides links to further information about where each tool should be used, and how
to use them.
2.5.1. t op
The top tool, provided by the procps-ng package, gives a dynamic view of the processes in a running
system. It can display a variety of information, including a system summary and a list of tasks
currently being managed by the Linux kernel. It also has a limited ability to manipulate processes,
and to make configuration changes persistent across system restarts.
By default, the processes displayed are ordered according to the percentage of CPU usage, so that
you can easily see the processes consuming the most resources. Both the information top displays
and its operation are highly configurable to allow you to concentrate on different usage statistics as
required.
For detailed information about using top, see the man page:
$ man top
2.5.2. ps
The ps tool, provided by the procps-ng package, takes a snapshot of a select group of active
processes. By default, the group examined is limited to processes that are owned by the current user
and associated with the terminal in which ps is run.
ps can provide more detailed information about processes than top, but by default it provides a
single snapshot of this data, ordered by process identifier.
For detailed information about using ps, see the man page:
$ man ps
2.5.3. Virt ual Memory St at ist ics (vmst at )

The Virtual Memory Statistics tool, vmstat, provides instant reports on your system's processes,
memory, paging, block input/output, interrupts, and CPU activity. Vmstat lets you set a sampling
interval so that you can observe system activity in near-real time.
vmstat is provided by the procps-ng package. For detailed information about using vmstat, see the
man page:
$ man vmstat
2.5.4 . Syst em Act ivit y Report er (sar)

The System Activity Reporter, sar, collects and reports information about system activity that has
occurred so far on the current day. The default output displays the current day's CPU usage at 10
minute intervals from the beginning of the day (00:00:00 according to your system clock).
You can also use the -i option to set the interval time in seconds, for example, sar -i 6 0 tells sar
to check CPU usage every minute.
sar is a useful alternative to manually creating periodic reports on system activity with top. It is
provided by the sysstat package. For detailed information about using sar, see the man page:
$ man sar
2.6. t uned and t uned-adm

tuned is a tuning daemon that can adapt the operating system to perform better under certain
workloads by setting a tuning profile. tuned-adm is a command line tool that lets users switch
between different tuning profiles.
Several pre-defined profiles are included for common use cases, but tuned-adm also lets you define
custom profiles, which can be either based on one of the pre-defined profiles, or defined from scratch.
In Red Hat Enterprise Linux 7 the default profile is thro ug hput-perfo rmance.
The profiles provided with tuned-adm are divided into two categories: power saving profiles, and
performance boosting profiles. The performance boosting profiles include profiles that focus on the
following:
low latency for storage and network

high throughput for storage and network
virtual machine performance
virtualization host performance
For detailed information about how to enabled tuned, see Section A.5, tuned .
For detailed information about the performance boosting profiles provided with tuned-adm, see
Section A.6, tuned-adm .
For detailed information about the power saving profiles provided with tuned-adm, see the Red Hat
Enterprise Linux 7 Power Management Guide, available from
For detailed information about using tuned and tuned-adm, see their respective man pages:
$ man tuned
$ man tuned-adm
2.7. perf
The perf tool uses hardware performance counters and kernel tracepoints to track the impact of other
commands and applications on your system. Various perf subcommands display and record
statistics for common performance events, and analyze and report on the data recorded.
For detailed information about perf and its subcommands, see Section A.7, perf .
Alternatively, more information is available in the Red Hat Enterprise Linux 7 Developer Guide,
available from http://access.redhat.com/site/documentation/Red_Hat_Enterprise_Linux/.
2.8. t urbost at
Turbostat is provided by the kernel-tools package. It reports on processor topology, frequency, idle
power-state statistics, temperature, and power usage on Intel 64 processors.
Turbostat is useful for identifying servers that are inefficient in terms of power usage or idle time. It
also helps to identify the rate of system management interrupts (SMIs) occurring on the system. It can
also be used to verify the effects of power management tuning.
Turbostat requires root privileges to run. It also requires processor support for the following:
invariant time stamp counters
APERF model-specific registers
MPERF model-specific registers
For more details about turbostat output and how to read it, see Section A.11, turbostat .
For more information about turbostat, see the man page:
$ man turbostat
2.9. iost at
The io st at tool is provided by the sysstat package. It monitors and reports on system input/output
device loading to help administrators make decisions about how to balance input/output load
between physical disks. It reports on processor or device utilization since iostat was last run, or since
boot. You can focus the output of these reports on specific devices by using the parameters defined
in the iostat man page:
$ man iostat
2.10. irqbalance
irq b alan ce is a command line tool that distributes hardware interrupts across processors to
improve system performance. For details about irq b alan ce, see Section A.1, irqbalance or the
man page:
$ man irqbalance
2.11. ss
ss is a command-line utility that prints statistical information about sockets, allowing administrators
to assess device performance over time. By default, ss lists open non-listening TCP sockets that have
established connections, but a number of useful options are provided to help administrators filter out
statistics about specific sockets.
Red Hat recommends using ss over netstat in Red Hat Enterprise Linux 7.
One common usage is ss -tmpi e which displays detailed information (including internal
information) about TCP sockets, memory usage, and processes using the socket.
ss is provided by the iproute package. For more information, see the man page:
$ man ss
2.12. numast at
The n u mast at tool displays memory statistics for processes and the operating system on a perNUMA-node basis.
By default, n u mast at displays per-node NUMA hit an miss system statistics from the kernel memory
allocator. Optimal performance is indicated by high numa_hi t values and low numa_mi ss values.
N u mast at also provides a number of command line options, which can show how system and
process memory is distributed across NUMA nodes in the system.
It can be useful to cross-reference per-node n u mast at output with per-CPU t o p output to verify that
process threads are running on the same node to which memory is allocated.
N u mast at is provided by the numactl package. For details about how to use numastat, see
Section A.12, numastat . For further information about numastat, see the man page:
$ man numastat
2.13. numad
numad is an automatic NUMA affinity management daemon. It monitors NUMA topology and resource
usage within a system in order to dynamically improve NUMA resource allocation and management
(and therefore system performance). D epending on system workload, numad can provide up to 50
percent improvements in performance benchmarks. It also provides a pre-placement advice service
that can be queried by various job management systems to provide assistance with the initial
binding of CPU and memory resources for their processes.
numad monitors available system resources on a per-node basis by periodically accessing
information in the /pro c file system. It tries to maintain a specified resource usage level, and
rebalances resource allocation when necessary by moving processes between NUMA nodes. numad
attempts to achieve optimal NUMA performance by localizing and isolating significant processes on
a subset of the system's NUMA nodes.
numad primarily benefits systems with long-running processes that consume significant amounts of
resources, and are contained in a subset of the total system resources. It may also benefit
applications that consume multiple NUMA nodes' worth of resources; however, the benefits provided
by numad decrease as the consumed percentage of system resources increases.
numad is unlikely to improve performance when processes run for only a few minutes, or do not
consume many resources. Systems with continuous, unpredictable memory access patterns, such as
large in-memory databases, are also unlikely to benefit from using numad.
For further information about using numad, see Section 3.3.5, Automatic NUMA affinity management
with numad or Section A.14, numad , or refer to the man page:
$ man numad
2.14 . Syst emT ap

SystemTap is a tracing and probing tool that lets you monitor and analyze operating system
activities, especially kernel activities, in fine detail. It provides information similar to the output of
tools like top, ps, netstat, and iostat, but includes additional options for filtering and analyzing
collected data.
SystemTap provides a deeper, more precise analysis of system activities and application behavior to
allow you to pinpoint system and application bottlenecks.
For more detailed information about SystemTap, see the Red Hat Enterprise Linux 7 SystemTap
Beginner's Guide and the Red Hat Enterprise Linux 7 SystemTap TapSet Reference. Both books are
2.15. OProfile
OProfile is a system-wide performance monitoring tool. It uses the processor's dedicated
performance monitoring hardware to retrieve information about the kernel and system executables to
determine the frequency of certain events, such as when memory is referenced, the number of secondlevel cache requests, and the number of hardware requests received. OProfile can also be used to
determine processor usage, and to determine which applications and services are used most often.
However, OProfile does have several limitations:
10
Performance monitoring samples may not be precise. Because the processor may execute
instructions out of order, samples can be recorded from a nearby instruction instead of the
instruction that triggered the interrupt.
OProfile expects processes to start and stop multiple times. As such, samples from multiple runs
are allowed to accumulate. You may need to clear the sample data from previous runs.
OProfile focuses on identifying problems with processes limited by CPU access. It is therefore not
useful for identifying processes that are sleeping while they wait for locks on other events.
For more detailed information about OProfile, see Section A.15, OProfile , or the Red Hat
Enterprise Linux 7 System Administrator's Guide, available from
http://access.redhat.com/site/documentation/Red_Hat_Enterprise_Linux/. Alternatively, refer to the
documentation on your system, located in /usr/share/d o c/o pro fi l e-version.
2.16. Valgrind
Valgrind provides a number of detection and profiling tools to help improve the performance of your
applications. These tools can detect memory and thread-related errors, as well as heap, stack, and
array overruns, letting you easily locate and correct errors in your application code. They can also
profile the cache, the heap, and branch-prediction to identify factors that may increase application
speed and minimize memory usage.
Valgrind analyzes your application by running it on a synthetic CPU and instrumenting existing
application code as it is executed. It then prints commentary that clearly identifies each process
involved in application execution to a user-specified file, file descriptor, or network socket. Note that
executing instrumented code can take between four and fifty times longer than normal execution.
Valgrind can be used on your application as-is, without recompiling. However, because Valgrind
uses debugging information to pinpoint issues in your code, if your application and support libraries
were not compiled with debugging information enabled, Red Hat recommends recompiling to include
this information.
Valgrind also integrates with the GNU Project D ebugger (gdb) to improve debugging efficiency.
Valgrind and its subordinate tools are useful for memory profiling. For detailed information about
using Valgrind to profile system memory, see Section 4.2.2, Profiling application memory usage with
Valgrind .
For detailed information about Valgrind, see the Red Hat Enterprise Linux 7 D eveloper Guide,
For detailed information about using Valgrind, see the man page:
$ man valgrind
Accompanying documentation can also be found in /usr/share/d o c/val g ri nd -version when
the valgrind package is installed.
11
Chapter 3. CPU
This chapter outlines CPU hardware details and configuration options that affect application
performance in Red Hat Enterprise Linux 7. Section 3.1, Considerations discusses the CPU related
factors that affect performance. Section 3.2, Monitoring and diagnosing performance problems
teaches you how to use Red Hat Enterprise Linux 7 tools to diagnose performance problems related
to CPU hardware or configuration details. Section 3.3, Configuration suggestions discusses the
tools and strategies you can use to solve CPU related performance problems in Red Hat
Enterprise Linux 7.
3.1. Considerat ions

Read this section to gain an understanding of how system and application performance is affected
by the following factors:
How processors are connected to each other and to related resources like memory.
How processors schedule threads for execution.
How processors handle interrupts in Red Hat Enterprise Linux 7.
3.1.1. Syst em T opology

In modern computing, the idea of a central processing unit is a misleading one, as most modern
systems have multiple processors. How these processors are connected to each other and to other
system resources the topology of the system can greatly affect system and application
performance, and the tuning considerations for a system.
There are two primary types of topology used in modern computing:
S ymmet ric Mu lt i- Pro cesso r ( SMP) t o p o lo g y
SMP topology allows all processors to access memory in the same amount of time.
However, because shared and equal memory access inherently forces serialized memory
accesses from all the CPUs, SMP system scaling constraints are now generally viewed as
unacceptable. For this reason, practically all modern server systems are NUMA machines.
N o n - U n if o rm Memo ry Access ( N U MA) t o p o lo g y
NUMA topology was developed more recently than SMP topology. In a NUMA system,
multiple processors are physically grouped on a socket. Each socket has a dedicated area
of memory, and processors that have local access to that memory are referred to collectively
as a node.
Processors on the same node have high speed access to that node's memory bank, and
slower access to memory banks not on their node. Therefore, there is a performance penalty
to accessing non-local memory.
Given this performance penalty, performance sensitive applications on a system with NUMA
topology should access memory that is on the same node as the processor executing the
application, and should avoid accessing remote memory wherever possible.
When tuning application performance on a system with NUMA topology, it is therefore
important to consider where the application is being executed, and which memory bank is
closest to the point of execution.
In a system with NUMA topology, the /sys file system contains information about how
12
Chapt er 3. CPU
processors, memory, and peripheral devices are connected. The

/sys/d evi ces/system/cpu directory contains details about how processors in the
system are connected to each other. The /sys/d evi ces/system/no d e directory
contains information about NUMA nodes in the system, and the relative distances between
those nodes.
3.1 .1 .1 . De t e rm ining syst e m t o po lo gy

There are a number of commands that can help you understand the topology of your system. The
numactl --hard ware command gives an overview of your system's topology.
$ numactl --hardware
available: 4 nodes (0-3)
node 0 cpus: 0 4 8 12 16 20 24 28 32 36
node 0 size: 65415 MB
node 0 free: 43971 MB
node 1 cpus: 2 6 10 14 18 22 26 30 34 38
node 2 cpus: 1 5 9 13 17 21 25 29 33 37
node 3 cpus: 3 7 11 15 19 23 27 31 35 39
node distances:
node
0
1
2
3
0: 10 21 21 21
1: 21 10 21 21
2: 21 21 10 21
3: 21 21 21 10
The l scpu command, provided by the util-linux package, gathers information about the CPU
architecture, such as the number of CPUs, threads, cores, sockets, and NUMA nodes.
$ lscpu
Architecture:
CPU op-mode(s):
Byte Order:
CPU(s):
On-line CPU(s) list:
Thread(s) per core:
Core(s) per socket:
Socket(s):
NUMA node(s):
Vendor ID:
CPU family:
Model:
Model name:
Stepping:
CPU MHz:
BogoMIPS:
Virtualization:
L1d cache:
L1i cache:
x86_64
32-bit, 64-bit
Little Endian
40
0-39
1
10
4
4
GenuineIntel
6
47
Intel(R) Xeon(R) CPU E7- 4870
2
2394.204
4787.85
VT-x
32K
32K
@ 2.40GHz
13
L2 cache:
L3 cache:
NUMA node0
NUMA node1
NUMA node2
NUMA node3
CPU(s):
CPU(s):
CPU(s):
CPU(s):
256K
30720K
0,4,8,12,16,20,24,28,32,36
2,6,10,14,18,22,26,30,34,38
1,5,9,13,17,21,25,29,33,37
3,7,11,15,19,23,27,31,35,39
The l sto po command, provided by the hwloc package, creates a graphical representation of your
system. The l sto po -no -g raphi cs command provides detailed textual output.
14
Chapt er 3. CPU
Output of lstopo command
15
3.1.2. Scheduling
In Red Hat Enterprise Linux, the smallest unit of process execution is called a thread. The system
scheduler determines which processor runs a thread, and for how long the thread runs. However,
because the scheduler's primary concern is to keep the system busy, it may not schedule threads
optimally for application performance.
For example, say an application on a NUMA system is running on Node A when a processor on
Node B becomes available. To keep the processor on Node B busy, the scheduler moves one of the
application's threads to Node B. However, the application thread still requires access to memory on
Node A. Because the thread is now running on Node B, and Node A memory is no longer local to the
thread, it will take longer to access. It may take longer for the thread to finish running on Node B than
it would have taken to wait for a processor on Node A to become available, and to execute the thread
on the original node with local memory access.
Performance sensitive applications often benefit from the designer or administrator determining
where threads are run. For details about how to ensure threads are scheduled appropriately for the
needs of performance sensitive applications, see Section 3.3.6, Tuning scheduling policy .
3.1 .2 .1 . Ke rne l T icks

In previous versions of Red Hat Enterprise Linux, the Linux kernel interrupted each CPU on a regular
basis to check what work needed to be done. It used the results to make decisions about process
scheduling and load balancing. This regular interruption was known as a kernel tick.
This tick occurred regardless of whether there was work for the core to do. This meant that even idle
cores were forced into higher power states on a regular basis (up to 1000 times per second) to
respond to the interrupts. This prevented the system from effectively using deep sleep states included
in recent generations of x86 processors.
In Red Hat Enterprise Linux 6 and 7, by default, the kernel no longer interrupts idle CPUs, which tend
to be in low power states. This behavior is known as the tickless kernel. Where one or fewer tasks are
running, periodic interrupts have been replaced with on-demand interrupts, allowing CPUs to remain
in an idle or low power state for longer, and reducing power usage.
Red Hat Enterprise Linux 7 offers a dynamic tickless option (no hz_ful l ) to further improve
determinism by reducing kernel interference with user-space tasks. This option can be enabled on
specified cores with the nohz_full kernel parameter. When this option is enabled on a core, all
timekeeping activities are moved to non-latency-sensitive cores. This can be useful for high
performance computing and realtime computing workloads where user-space tasks are particularly
sensitive to microsecond-level latencies associated with the kernel timer tick.
For details on how to enable the dynamic tickless behavior in Red Hat Enterprise Linux 7, see
Section 3.3.1, Configuring kernel tick time .
3.1.3. Int errupt Request (IRQ) Handling

An interrupt request or IRQ is a signal for immediate attention sent from a piece of hardware to a
processor. Each device in a system is assigned one or more IRQ numbers to allow it to send unique
interrupts. When interrupts are enabled, a processor that receives an interrupt request will
immediately pause execution of the current application thread in order to address the interrupt
request.
Because they halt normal operation, high interrupt rates can severely degrade system performance. It
is possible to reduce the amount of time taken by interrupts by configuring interrupt affinity or by
sending a number of lower priority interrupts in a batch (coalescing a number of interrupts).
16
Chapt er 3. CPU
For more information about tuning interrupt requests, see Section 3.3.7, Setting interrupt affinity or
Section 3.3.8, Configuring CPU, thread, and interrupt affinity with Tuna . For information specific to
network interrupts, see Chapter 6, Networking.
3.2. Monit oring and diagnosing performance problems

Red Hat Enterprise Linux 7 provides a number of tools that are useful for monitoring system
performance and diagnosing performance problems related to processors and their configuration.
This section outlines the available tools and gives examples of how to use them to monitor and
diagnose processor related performance issues.
3.2.1. t urbost at
T u rb o st at prints counter results at specified intervals to help administrators identify unexpected
behavior in servers, such as excessive power usage, failure to enter deep sleep states, or system
management interrupts (SMIs) being created unnecessarily.
The t u rb o st at tool is part of the kernel-tools package. It is supported for use on systems with AMD 64
and Intel 64 processors. It requires root privileges to run, and processor support for invariant time
stamp counters, and APERF and MPERF model specific registers.
For usage examples, see the man page:
$ man turbostat
3.2.2. numast at
Important
This tool received substantial updates in the Red Hat Enterprise Linux 6 life cycle. While the
default output remains compatible with the original tool written by Andi Kleen, supplying any
options or parameters to numastat significantly changes the format of its output.
The n u mast at tool displays per-NUMA node memory statistics for processes and the operating
system and shows administrators whether process memory is spread throughout a system or
centralized on specific nodes.
Cross reference n u mast at output with per-processor t o p output to confirm that process threads are
running on the same node from which process memory is allocated.
N u mast at is provided by the numactl package. For further information about n u mast at output, see
the man page:
$ man numastat
3.2.3. /proc/int errupt s

The /pro c/i nterrupts file lists the number of interrupts sent to each processor from a particular
I/O device. It displays the interrupt request (IRQ) number, the number of that type of interrupt request
handled by each processor in the system, the type of interrupt sent, and a comma-separated list of
devices that respond to the listed interrupt request.
17
If a particular application or device is generating a large number of interrupt requests to be handled

by a remote processor, its performance is likely to suffer. In this case, poor performance can be
alleviated by having a processor on the same node as the application or device handle the interrupt
requests. For details on how to assign interrupt handling to a specific processor, see Section 3.3.7,
Setting interrupt affinity .
3.3. Configurat ion suggest ions

Red Hat Enterprise Linux provides a number of tools to assist administrators in configuring the
system. This section outlines the available tools and provides examples of how they can be used to
solve processor related performance problems in Red Hat Enterprise Linux 7.
3.3.1. Configuring kernel t ick t ime

By default, Red Hat Enterprise Linux 7 uses a tickless kernel, which does not interrupt idle CPUs in
order to reduce power usage and allow newer processors to take advantage of deep sleep states.
Red Hat Enterprise Linux 7 also offers a dynamic tickless option (disabled by default), which is useful
for very latency-sensitive workloads, such as high performance computing or realtime computing.
To enable dynamic tickless behavior in certain cores, specify those cores on the kernel command
line with the nohz_full parameter. On a 16 core system, specifying no hz_ful l = 1-15 enables
dynamic tickless behavior on cores 1 through 15, moving all timekeeping to the only unspecified core
(core 0). This behavior can be enabled either temporarily at boot time, or persistently in the
/etc/d efaul t/g rub file. For persistent behavior, run the g rub2-mkco nfi g -o
/bo o t/g rub2/g rub. cfg command to save your configuration.
Enabling dynamic tickless behavior does require some manual administration.
When the system boots, you must manually move rcu threads to the non-latency-sensitive core, in
this case core 0.
# for i in `pgrep rcu[^c]` ; do taskset -pc 0 $i ; done
Use the isolcpus parameter on the kernel command line to isolate certain cores from user-space
tasks.
Optionally, set CPU affinity for the kernel's write-back bdi-flush threads to the housekeeping core:
echo 1 > /sys/bus/workqueue/devices/writeback/cpumask
Verify that the dynamic tickless configuration is working correctly by executing the following
command, where stress is a program that spins on the CPU for 1 second.
# perf stat -C 1 -e irq_vectors:local_timer_entry taskset -c 1 stress -t
1 -c 1
One possible replacement for stress is a script that runs something like whi l e : ; d o d = 1;
d o ne. The program available at the following link is another suitable replacement:
https://dl.fedoraproject.org/pub/epel/6/x86_64/repoview/stress.html.
The default kernel timer configuration shows 1000 ticks on a busy CPU:
18
Chapt er 3. CPU

1 -c 1
1000 irq_vectors:local_timer_entry
With the dynamic tickless kernel configured, you should see 1 tick instead:
1 -c 1
1 irq_vectors:local_timer_entry
3.3.2. Set t ing hardware performance policy (x86_energy_perf_policy)

The x86 _en erg y_p erf _p o licy tool allows administrators to define the relative importance of
performance and energy efficiency. This information can then be used to influence processors that
support this feature when they select options that trade off between performance and energy
efficiency.
By default, it operates on all processors in perfo rmance mode. It requires processor support, which
is indicated by the presence of C P UID . 0 6 H. EC X. bi t3, and must be run with root privileges.
x86 _en erg y_p erf _p o licy is provided by the kernel-tools package. For details of how to use
x86 _en erg y_p erf _p o licy, see Section A.10, x86_energy_perf_policy or refer to the man page:
$ man x86_energy_perf_policy
3.3.3. Set t ing process affinit y wit h t askset

The t askset tool is provided by the util-linux package. T askset allows administrators to retrieve and
set the processor affinity of a running process, or launch a process with a specified processor
affinity.
Important
t askset does not guarantee local memory allocation. If you require the additional
performance benefits of local memory allocation, Red Hat recommends using n u mact l
instead of t askset .
For more information about t askset , see Section A.16, taskset or the man page:
$ man taskset
3.3.4 . Managing NUMA affinit y wit h numact l

Administrators can use n u mact l to run a process with a specified scheduling or memory placement
policy. N u mact l can also set a persistent policy for shared memory segments or files, and set the
processor affinity and memory affinity of a process.
In a system with NUMA topology, a processor's memory access slows as the distance between the
processor and the memory bank increases. Therefore, it is important to configure applications that
are sensitive to performance so that they allocate memory from the closest possible memory bank. It
is best to use memory and CPUs that are in the same NUMA node.
19
Multi-threaded applications that are sensitive to performance may benefit from being configured to
execute on a specific NUMA node rather than a specific processor. Whether this is suitable depends
on your system and the requirements of your application. If multiple application threads access the
same cached data, then configuring those threads to execute on the same processor may be
suitable. However, if multiple threads that access and cache different data execute on the same
processor, each thread may evict cached data accessed by a previous thread. This means that each
thread 'misses' the cache, and wastes execution time fetching data from disk and replacing it in the
cache. You can use the p erf tool, as documented in Section A.7, perf , to check for an excessive
number of cache misses.
N u mact l provides a number of options to assist you in managing processor and memory affinity.
See Section A.12, numastat or the man page for details:
$ man numactl
Note
The n u mact l package includes the l i bnuma library. This library offers a simple
programming interface to the NUMA policy supported by the kernel, and can be used for more
fine-grained tuning than the n u mact l application. For more information, see the man page:
$ man numa
3.3.5. Aut omat ic NUMA affinit y management wit h numad

numad is an automatic NUMA affinity management daemon. It monitors NUMA topology and resource
usage within a system in order to dynamically improve NUMA resource allocation and management.
n u mad also provides a pre-placement advice service that can be queried by various job
management systems to provide assistance with the initial binding of CPU and memory resources for
their processes. This pre-placement advice is available regardless of whether numad is running as
an executable or a service.
For details of how to use n u mad , see Section A.14, numad or refer to the man page:
$ man numad
3.3.6. T uning scheduling policy

The Linux scheduler implements a number of scheduling policies, which determine where and for
how long a thread runs. There are two major categories of scheduling policies: normal policies and
realtime policies. Normal threads are used for tasks of normal priority. Realtime policies are used for
time-sensitive tasks that must complete without interruptions.
Realtime threads are not subject to time slicing. This means they will run until they block, exit,
voluntarily yield, or are pre-empted by a higher priority thread. The lowest priority realtime thread is
scheduled before any thread with a normal policy.
3.3.6 .1 . Sche duling po licie s

3.3.6 .1.1. St at ic p rio rit y sch ed u lin g wit h SC H ED _FIFO
20
Chapt er 3. CPU
SC HED _FIFO (also called static priority scheduling) is a realtime policy that defines a fixed priority
for each thread. This policy allows administrators to improve event response time and reduce
latency, and is recommended for time sensitive tasks that do not run for an extended period of time.
When SC HED _FIFO is in use, the scheduler scans the list of all SC HED _FIFO threads in priority
order and schedules the highest priority thread that is ready to run. The priority level of a
SC HED _FIFO thread can be any integer from 1 to 99, with 99 treated as the highest priority. Red Hat
recommends starting at a low number and increasing priority only when you identify latency issues.
Warning
Because realtime threads are not subject to time slicing, Red Hat does not recommend setting
a priority of 99. This places your process at the same priority level as migration and watchdog
threads; if your thread goes into a computational loop and these threads are blocked, they will
not be able to run. Systems with a single processor will eventually hang in this situation.
Administrators can limit SC HED _FIFO bandwidth to prevent realtime application programmers from
initiating realtime tasks that monopolize the processor.
/p ro c/sys/kern el/sch ed _rt _p erio d _u s
This parameter defines the time period in microseconds that is considered to be one
hundred percent of processor bandwidth. The default value is 10 0 0 0 0 0 s, or 1 second.
/p ro c/sys/kern el/sch ed _rt _ru n t ime_u s
This parameter defines the time period in microseconds that is devoted to running realtime
threads. The default value is 9 50 0 0 0 s, or 0.95 seconds.
3.3.6 .1.2. R o u n d ro b in p rio rit y sch ed u lin g wit h SC H ED _R R
SC HED _R R is a round-robin variant of SC HED _FIFO . This policy is useful when multiple threads
need to run at the same priority level.
Like SC HED _FIFO , SC HED _R R is a realtime policy that defines a fixed priority for each thread. The
scheduler scans the list of all SC HED _R R threads in priority order and schedules the highest priority
thread that is ready to run. However, unlike SC HED _FIFO , threads that have the same priority are
scheduled round-robin style within a certain time slice.
You can set the value of this time slice in milliseconds with the sched_rr_timeslice_ms kernel
parameter (/pro c/sys/kernel /sched _rr_ti mesl i ce_ms). The lowest value is 1 millisecond.
3.3.6 .1.3. N o rmal sch ed u lin g wit h SC H ED _O T H ER
SC HED _O T HER is the default scheduling policy in Red Hat Enterprise Linux 7. This policy uses the
Completely Fair Scheduler (CFS) to allow fair processor access to all threads scheduled with this
policy. This policy is most useful when there are a large number of threads or data throughput is a
priority, as it allows more efficient scheduling of threads over time.
When this policy is in use, the scheduler creates a dynamic priority list based partly on the niceness
value of each process thread. Administrators can change the niceness value of a process, but
cannot change the scheduler's dynamic priority list directly.
For details about changing process niceness, see the Red Hat Enterprise Linux 7 Deployment Guide,
21
3.3.6 .2 . Iso lat ing CPUs

You can isolate one or more CPUs from the scheduler with the isolcpus boot parameter. This
prevents the scheduler from scheduling any user-space threads on this CPU.
Once a CPU is isolated, you must manually assign processes to the isolated CPU, either with the
CPU affinity system calls or the numactl command.
To isolate the third and sixth to eighth CPUs on your system, add the following to the kernel
command line:
isolcpus=2,5-7
You can also use the T u n a tool to isolate a CPU. T u n a can isolate a CPU at any time, not just at
boot time. However, this method of isolation is subtly different from the isolcpus parameter, and
does not currently achieve the performance gains associated with isolcpus. See Section 3.3.8,
Configuring CPU, thread, and interrupt affinity with Tuna for more details about this tool.
3.3.7. Set t ing int errupt affinit y

Interrupt requests have an associated affinity property, smp_affinity, that defines the processors
that will handle the interrupt request. To improve application performance, assign interrupt affinity
and process affinity to the same processor, or processors on the same core. This allows the specified
interrupt and application threads to share cache lines.
The interrupt affinity value for a particular interrupt request is stored in the associated
/pro c/i rq /irq_number/smp_affi ni ty file. smp_affinity is stored as a hexadecimal bit
mask representing all processors in the system. The default value is f, meaning that an interrupt
request can be handled on any processor in the system. Setting this value to 1 means that only
processor 0 can handle the interrupt.
On systems with more than 32 processors, you must delimit smp_affinity values for discrete 32 bit
groups. For example, if you want only the first 32 processors of a 64 processor system to service an
interrupt request, you could run:
# echo 0xffffffff,00000000 > /proc/irq/IRQ_NUMBER/smp_affinity
Alternatively, if the BIOS exports its NUMA topology, the irq b alan ce service can use that information
to serve interrupt requests on the node that is local to the hardware requesting service. For details
about irq b alan ce, see Section A.1, irqbalance
Note
On systems that support interrupt steering, modifying the smp_affinity of an interrupt
request sets up the hardware so that the decision to service an interrupt with a particular
processor is made at the hardware level with no intervention from the kernel. For more
information about interrupt steering, see Chapter 6, Networking.
3.3.8. Configuring CPU, t hread, and int errupt affinit y wit h T una
T u n a can control CPU, thread, and interrupt affinity, and provides a number of actions for each type
of entity it can control. For a full list of T u n a's capabilities, see Section A.2, Tuna .
22
Chapt er 3. CPU
To move all threads away from one or more specified CPUs, run the following command, replacing
CPUs with the number of the CPU you want to isolate.
# tuna --cpus CPUs --isolate
To include a CPU in the list of CPUs that can run certain threads, run the following command,
replacing CPUs with the number of the CPU you want to include.
# tuna --cpus CPUs --include
To move an interrupt request to a specified CPU, run the following command, replacing CPU with the
number of the CPU, and IRQs with the comma-delimited list of interrupt requests you want to move.
# tuna --irqs IRQs --cpus CPU --move
Alternatively, you can use the following command to find all interrupt requests of the pattern sfc1*.
# tuna -q sfc1* -c7 -m -x
To change the policy and priority of a thread, run the following command, replacing thread with the
thread you want to change, policy with the name of the policy you want the thread to operate under,
and level with an integer from 0 (lowest priority) to 99 (highest priority).
# tuna --threads thread --priority policy:level
23
Chapter 4. Memory
This chapter outlines the memory management capabilities of Red Hat Enterprise Linux 7. Section 4.1,
Considerations discusses memory related factors that affect performance. Section 4.2, Monitoring
and diagnosing performance problems teaches you how to use Red Hat Enterprise Linux 7 tools to
diagnose performance problems related to memory utilization or configuration details. Section 4.3,
Configuration tools discusses the tools and strategies you can use to solve memory related
performance problems in Red Hat Enterprise Linux 7.
4 .1. Considerat ions

By default, Red Hat Enterprise Linux 7 is optimized for moderate workloads. If your application or use
case requires a large amount of memory, changing the way your system manages virtual memory
may improve the performance of your application.
4 .1.1. Page siz e

Physical memory is managed in chunks called pages. The physical location of each page is mapped
to a virtual location so that the processor can access the memory. This mapping is stored in a data
structure known as the page table.
By default, a page is about 4 KB in size. Since the default page size is so small, you need an
enormous number of pages to manage a large amount of memory. However, the page table can only
store a finite number of address mappings, and increasing the number of address mappings it can
store is both expensive and difficult in terms of maintaining performance levels as memory
requirements scale.
Red Hat Enterprise Linux also offers the ability to manage a larger amount of memory per page with
static huge pages. Static huge pages can be configured up to sizes of 1 GB. However, they are
difficult to manage manually, and must be assigned at boot time. As of Red Hat Enterprise Linux 7.1,
they can be assigned on a per-node basis.
Transparent huge pages are a largely automated alternative to static huge pages. Transparent huge
pages are 2 MB in size, and are enabled by default. They can sometimes interfere with latencysensitive applications, so are often disabled when latency is a concern.
For details about configuring huge pages to improve application performance, see Section 4.3.1,
Configuring huge pages .
4 .1.2. T ranslat ion Lookaside Buffer siz e

Reading address mappings from the page table is time consuming and resource expensive, so the
Linux operating system provides a cache for recently-used addresses: the Translation Lookaside
Buffer (TLB). However, the default TLB can only cache a certain number of address mappings. If a
requested address mapping is not in the TLB (that is, the TLB is missed), the system still needs to
read the page table to determine the physical to virtual address mapping.
Because of the relationship between application memory requirements and the size of pages used to
cache address mappings, applications with large memory requirements are more likely to suffer
performance degradation from TLB misses than applications with minimal memory requirements. It is
therefore important to avoid TLB misses wherever possible.
24
Chapt er 4 . Memory
Red Hat Enterprise Linux provides the Huge Translation Lookaside Buffer (HugeTLB), which allows
memory to be managed in very large segments. This lets a greater number of address mappings be
cached at one time, which reduces the likelihood of TLB misses, thereby improving performance in
applications with large memory requirements.
For details about configuring HugeTLB, see Section 4.3.1, Configuring huge pages .
4 .2. Monit oring and diagnosing performance problems

performance and diagnosing performance problems related to system memory. This section outlines
the available tools and gives examples of how to use them to monitor and diagnose memory related
performance issues.
4 .2.1. Monit oring memory usage wit h vmst at

Vmst at , provided by the procps-ng package, outputs reports on your system's processes, memory,
paging, block input/output, interrupts, and CPU activity. It provides an instantaneous report of the
average of these events since the machine was last booted, or since the previous report.
The following command displays a table of various event counters and memory statistics.
$ vmstat -s
For further details of how to use vmst at , see Section A.9, vmstat or the man page:
$ man vmstat
4 .2.2. Profiling applicat ion memory usage wit h Valgrind

Valg rin d is a framework that provides instrumentation to user-space binaries. It ships with a number
of tools that can be used to profile and analyze program performance. The valg rin d tools outlined
in this section can help you to detect memory errors such as uninitialized memory use and improper
memory allocation or deallocation.
To use valg rin d or any of its tools, install the valgrind package:
# yum install valgrind
4 .2 .2 .1 . Pro filing m e m o ry usage wit h Me m che ck

Memch eck is the default valg rin d tool. It detects and reports on a number of memory errors that
can be difficult to detect and diagnose, such as:
Memory access that should not occur
Undefined or uninitialized value use
Incorrectly freed heap memory
Pointer overlap
Memory leaks
25
Note
Memch eck can only report these errors; it cannot prevent them from occurring. If your
program accesses memory in a way that would normally cause a segmentation fault, the
segmentation fault still occurs. However, memch eck will log an error message immediately
prior to the fault.
Because memch eck uses instrumentation, applications executed with memch eck run ten to thirty
times slower than usual.
To run memch eck on an application, execute the following command:
# valgrind --tool=memcheck application
You can also use the following options to focus memch eck output on specific types of problem.
- - leak- ch eck
After the application finishes executing, memch eck searches for memory leaks. The default
value is --l eak-check= summary, which prints the number of memory leaks found. You
can specify --l eak-check= yes or --l eak-check= ful l to output details of each
individual leak. To disable, specify --l eak-check= no .
- - u n d ef - valu e- erro rs
The default value is --und ef-val ue-erro rs= yes, which reports errors when undefined
values are used. You can also specify --und ef-val ue-erro rs= no , which will disable
this report and slightly speed up Memcheck.
- - ig n o re- ran g es
Specifies one or more ranges that memch eck should ignore when checking for memory
addressability, for example, --i g no re-rang es= 0 xP P -0 xQ Q ,0 xR R -0 xSS.
For a full list of memch eck options, see the documentation included at
/usr/share/d o c/val g ri nd -versi o n/val g ri nd _manual . pd f.
4 .2 .2 .2 . Pro filing cache usage wit h Cache grind

C ach eg rin d simulates application interaction with a system's cache hierarchy and branch
predictor. It tracks usage of the simulated first level instruction and data caches to detect poor code
interaction with this level of cache. It also tracks the last level of cache (second or third level) in order
to track memory access. As such, applications executed with cach eg rin d run twenty to one hundred
times slower than usual.
C ach eg rin d gathers statistics for the duration of application execution and outputs a summary to
the console. To run cach eg rin d on an application, execute the following command:
# valgrind --tool=cachegrind application
You can also use the following options to focus cach eg rin d output on a specific problem.
- - I1
Specifies the size, associativity, and line size of the first level instruction cache, like so: -I1= size,associativity,line_size.
26
Chapt er 4 . Memory
- - D 1
Specifies the size, associativity, and line size of the first level data cache, like so: -D 1= size,associativity,line_size.
- - LL
Specifies the size, associativity, and line size of the last level cache, like so: -LL= size,associativity,line_size.
- - cach e- sim
Enables or disables the collection of cache access and miss counts. This is enabled (-cache-si m= yes) by default. D isabling both this and --branch-si m leaves cach eg rin d
with no information to collect.
- - b ran ch - sim
Enables or disables the collection of branch instruction and incorrect prediction counts.
This is enabled (--branch-si m= yes) by default. D isabling both this and --cache-si m
leaves cach eg rin d with no information to collect.
C ach eg rin d writes detailed profiling information to a per-process cacheg ri nd . o ut. pid
file, where pid is the process identifier. This detailed information can be further processed by
the companion cg _an n o t at e tool, like so:
# cg_annotate cachegrind.out.pid
C ach eg rin d also provides the cg_diff tool, which makes it easier to chart program performance
before and after a code change. To compare output files, execute the following command, replacing
first with the initial profile output file, and second with the subsequent profile output file.
# cg_diff first second
The resulting output file can be viewed in greater detail with the cg _an n o t at e tool.
For a full list of cach eg rin d options, see the documentation included at
/usr/share/d o c/val g ri nd -versi o n/val g ri nd _manual . pd f.
4 .2 .2 .3. Pro filing he ap and st ack space wit h Massif

Massif measures the heap space used by a specified application. It measures both useful space
and any additional space allocated for bookkeeping and alignment purposes. massif helps you
understand how you can reduce your application's memory use to increase execution speed and
reduce the likelihood that your application will exhaust system swap space. Applications executed
with massif run about twenty times slower than usual.
To run massif on an application, execute the following command:
# valgrind --tool=massif application
You can also use the following options to focus massif output on a specific problem.
- - h eap
Specifies whether massif profiles the heap. The default value is --heap= yes. Heap
profiling can be disabled by setting this to --heap= no .
27
- - h eap - ad min
Specifies the number of bytes per block to use for administration when heap profiling is
enabled. The default value is 8 bytes.
- - st acks
Specifies whether massif profiles the stack. The default value is --stack= no , as stack
profiling can greatly slow massif . Set this option to --stack= yes to enable stack
profiling. Note that massif assumes that the main stack starts with a size of zero in order to
better indicate the changes in stack size that relate to the application being profiled.
- - t ime- u n it
Specifies the interval at which massif gathers profiling data. The default value is i
(instructions executed). You can also specify ms (milliseconds, or realtime) and B (bytes
allocated or deallocated on the heap and stack). Examining bytes allocated is useful for
short run applications and for testing purposes, as it is most reproducible across different
hardware.
Massif outputs profiling data to a massi f. o ut. pid file, where pid is the process identifier of the
specified application. The ms_p rin t tool graphs this profiling data to show memory consumption
over the execution of the application, as well as detailed information about the sites responsible for
allocation at points of peak memory allocation. To graph the data from the massi f. o ut. pid file,
execute the following command:
# ms_print massif.out.pid
For a full list of Massif options, see the documentation included at /usr/share/d o c/val g ri nd versi o n/val g ri nd _manual . pd f.
4 .3. Configurat ion t ools

Memory usage is typically configured by setting the value of one or more kernel parameters. These
parameters can be set temporarily by altering the contents of files in the /proc file system, or they can
be set persistently with the sysctl tool, which is provided by the procps-ng package.
For example, to set the overcommit_memory parameter to 1 temporarily, run the following command:
# echo 1 > /proc/sys/vm/overcommit_memory
To set this value persistently, add sysctl vm.overcommit_memory=1 in /etc/sysctl . co nf then run
the following command:
# sysctl -p
Setting a parameter temporarily is useful for determining the effect the parameter has on your system.
You can then set the parameter persistently when you are sure that the parameter's value has the
desired effect.
4 .3.1. Configuring huge pages

Huge pages rely on contiguous areas of memory, so it is best to define huge pages at boot time,
before memory becomes fragmented. To do so, add the following parameters to the kernel boot
command line:
28
Chapt er 4 . Memory
h u g ep ag es
D efines the number of persistent huge pages configured in the kernel at boot time. The
default value is 0. It is only possible to allocate (or deallocate) huge pages if there are
sufficient physically contiguous free pages in the system. Pages reserved by this parameter
cannot be used for other purposes.
This value can be adjusted after boot by changing the value of the
/pro c/sys/vm/nr_hug epag es file.
In a NUMA system, huge pages assigned with this parameter are divided equally between
nodes. You can assign huge pages to specific nodes at runtime by changing the value of
the node's /sys/d evi ces/system/no d e/node_id/hug epag es/hug epag es10 4 8576 kB/nr_hug epag es file.
For more information, read the relevant kernel documentation, which is installed in
/usr/share/d o c/kernel -d o ckernel _versi o n/D o cumentati o n/vm/hug etl bpag e. txt by default.
h u g ep ag esz
D efines the size of persistent huge pages configured in the kernel at boot time. Valid values
are 2 MB and 1 GB. The default value is 2 MB.
d ef au lt _h u g ep ag esz
D efines the default size of persistent huge pages configured in the kernel at boot time. Valid
values are 2 MB and 1 GB. The default value is 2 MB.
You can also use the following parameters to influence huge page behaviour during runtime.
/sys/d evices/syst em/n o d e/node_id/h u g ep ag es/h u g ep ag es- size/n r_h u g ep ag es
D efines the number of huge pages of the specified size assigned to the specified NUMA
node. This is supported as of Red Hat Enterprise Linux 7.1. The following example moves
adds twenty 2048 kB huge pages to no d e2.
# numastat -cm | egrep 'Node|Huge'
Node 0 Node 1 Node 2 Node 3
Total add
sysctl vm.overcommit_memory=1
in /etc/sysctl.conf
and execute
AnonHugePages
0
2
0
8
10
HugePages_Total
0
0
0
0
0
HugePages_Free
0
0
0
0
0
HugePages_Surp
0
0
0
0
0
# echo 20 > /sys/devices/system/node/node2/hugepages/hugepages2048kB/nr_hugepages
# numastat -cm | egrep 'Node|Huge'
Node 0 Node 1 Node 2 Node 3 Total
AnonHugePages
0
2
0
8
10
HugePages_Total
0
0
40
0
40
HugePages_Free
0
0
40
0
40
HugePages_Surp
0
0
0
0
0
29
/p ro c/sys/vm/n r_o verco mmit _h u g ep ag es

D efines the maximum number of additional huge pages that can be created and used by
the system through overcommitting memory. Writing any non-zero value into this file
indicates that the system obtains that number of huge pages from the kernel's normal page
pool if the persistent huge page pool is exhausted. As these surplus huge pages become
unused, they are then freed and returned to the kernel's normal page pool.
4 .3.2. Configuring syst em memory capacit y

This section discusses memory-related kernel parameters that may be useful in improving memory
utilization on your system. These parameters can be temporarily set for testing purposes by altering
the value of the corresponding file in the /pro c file system. Once you have determined the values
that produce optimal performance for your use case, you can set them permanently by using the
sysctl command.
4 .3.2 .1 . Virt ual Me m o ry param e t e rs

The parameters listed in this section are located in /pro c/sys/vm unless otherwise indicated.
d irt y_rat io
A percentage value. When this percentage of total system memory is modified, the system
begins writing the modifications to disk with the pd fl ush operation. The default value is
20 percent.
d irt y_b ackg ro u n d _rat io
A percentage value. When this percentage of total system memory is modified, the system
begins writing the modifications to disk in the background. The default value is 10 percent.
o verco mmit _memo ry
D efines the conditions that determine whether a large memory request is accepted or
denied.
The default value is 0 . By default, the kernel performs heuristic memory overcommit
handling by estimating the amount of memory available and failing requests that are too
large. However, since memory is allocated using a heuristic rather than a precise algorithm,
overloading memory is possible with this setting.
When this parameter is set to 1, the kernel performs no memory overcommit handling. This
increases the possibility of memory overload, but improves performance for memoryintensive tasks.
When this parameter is set to 2, the kernel denies requests for memory equal to or larger
than the sum of total available swap space and the percentage of physical RAM specified in
overcommit_ratio. This reduces the risk of overcommitting memory, but is recommended
only for systems with swap areas larger than their physical memory.
o verco mmit _rat io
Specifies the percentage o f physical RAM considered when overcommit_memory is set to
2. The default value is 50 .
max_map _co u n t
30
Chapt er 4 . Memory
D efines the maximum number of memory map areas that a process can use. The default
value (6 5530 ) is appropriate for most cases. Increase this value if your application needs
to map more than this number of files.
min _f ree_kb yt es
Specifies the minimum number of kilobytes to keep free across the system. This is used to
determine an appropriate value for each low memory zone, each of which is assigned a
number of reserved free pages in proportion to their size.
Warning
Extreme values can damage your system. Setting min_free_kbytes to an
extremely low value prevents the system from reclaiming memory, which can result in
system hangs and OOM-killing processes. However, setting min_free_kbytes too
high (for example, to 510% of total system memory) causes the system to enter an
out-of-memory state immediately, resulting in the system spending too much time
reclaiming memory.
o o m_ad j
In the event that the system runs out of memory and the panic_on_oom parameter is set to
0 , the o o m_ki l l er function kills processes until the system can recover, starting from the
process with the highest o o m_sco re.
The oom_adj parameter helps determine the oom_score of a process. This parameter is
set per process identifier. A value of -17 disables the o o m_ki l l er for that process. Other
valid values are from -16 to 15.
Note
Processes spawned by an adjusted process inherit that process's o o m_sco re.
s wap p in ess
A value from 0 to 10 0 which controls the degree to which the system favour anonymous
memory or the page cache. A high value improves file-system performance, while
aggressively swapping less active processes out of RAM. A low value avoids swapping
processes out of memory, which usually decrease latency, at the cost of I/O performance.
The default value is 6 0 .
Warning
Setting swappi ness= = 0 will very aggressively avoids swapping out, which
increase the risk of OOM killing under strong memory and I/O pressure.
4 .3.2 .2 . File syst e m param e t e rs

Parameters listed in this section are located in /pro c/sys/fs unless otherwise indicated.
a io - max- n r
31
D efines the maximum allowed number of events in all active asynchronous input/output
contexts. The default value is 6 5536 . Modifying this value does not pre-allocate or resize
any kernel data structures.
f ile- max
D efines the maximum number of file handles allocated by the kernel. The default value
matches the value of fi l es_stat. max_fi l es in the kernel, which is set to the largest
value out of either NR_FILE (8192 in Red Hat Enterprise Linux), or the result of the
following:
(mempages * (PAGE_SIZE / 1024)) / 10
Raising this value can resolve errors caused by a lack of available file handles.
4 .3.2 .3. Ke rne l param e t e rs

Parameters listed in this section are located under /pro c/sys/kernel unless otherwise indicated.
msg max
D efines the maximum allowable size in bytes of any single message in a message queue.
This value must not exceed the size of the queue (msgmnb). The default value is 6 5536 .
msg mn b
D efines the maximum size in bytes of a single message queue. The default value is 6 5536 .
msg mn i
D efines the maximum number of message queue identifiers (and therefore the maximum
number of queues). The default value on systems with 64-bit architecture is 19 85.
s h mall
D efines the total amount of shared memory in pages that can be used on the system at one
time. As of Red Hat Enterprise Linux, the default value is 1<<24 , or 33554432 pages.
s h mmax
D efines the maximum size of a single shared memory segment allowed by the kernel, in
pages. As of Red Hat Enterprise Linux, the default value is 1<<24 , or 33554432 bytes.
s h mmn i
D efines the system-wide maximum number of shared memory segments. The default value is
4 0 9 6 on all systems.
t h read s- max
D efines the system-wide maximum number of threads available to the kernel at one time.
The default value is equal to the value of the kernel parameter max_threads, or the result
of:
mempages / (8 * THREAD_SIZE / PAGE SIZE )
The minimum value is 20 .
32
Chapt er 5. St orage and File Syst ems
Chapter 5. Storage and File Systems

This chapter outlines supported file systems and configuration options that affect application
performance for both I/O and file systems in Red Hat Enterprise Linux 7. Section 5.1,
Considerations discusses the I/O and file system related factors that affect performance.
Section 5.2, Monitoring and diagnosing performance problems teaches you how to use Red Hat
Enterprise Linux 7 tools to diagnose performance problems related to I/O or file system configuration
details. Section 5.3, Configuration tools discusses the tools and strategies you can use to solve I/O
and file system related performance problems in Red Hat Enterprise Linux 7.

The appropriate settings for storage and file system performance are highly dependent on the
purpose of the storage. I/O and file system performance can be affected by any of the following
factors:
D ata write or read patterns
D ata alignment with underlying geometry
Block size
File system size
Journal size and location
Recording access times
Ensuring data reliability
Pre-fetching data
Pre-allocating disk space
File fragmentation
Resource contention
Read this chapter to gain an understanding of the formatting and mount options that affect file
system throughput, scalability, responsiveness, resource usage, and availability.
5.1.1. Solid-St at e Disks

Solid-state disks (SSD ) use NAND flash chips rather than rotating magnetic platters to store
persistent data. They provide a constant access time for data across their full Logical Block Address
range, and do not incur measurable seek costs like their rotating counterparts. They are more
expensive per gigabyte of storage space and have a lesser storage density, but they also have lower
latency and greater throughput than HD D s.
Performance generally degrades as the used blocks on an SSD approach the capacity of the disk.
The degree of degradation varies by vendor, but all devices experience degradation in this
circumstance. Enabling discard behavior can help to alleviate this degradation; see Section 5.1.4.3,
Maintenance for details.
The default I/O scheduler and virtual memory options are suitable for use with SSD s.
5.1.2. I/O Schedulers
33
The I/O scheduler determines when and for how long I/O operations run on a storage device. It is
also known as the I/O elevator.
Red Hat Enterprise Linux 7 provides three I/O schedulers.
d ead lin e
The default I/O scheduler for all block devices except SATA disks. D ead l i ne attempts to
provide a guaranteed latency for requests from the point at which requests reach the I/O
scheduler. This scheduler is suitable for most use cases, but particularly those in which
read operations occur more often than write operations.
Queued I/O requests are sorted into a read or write batch and then scheduled for execution
in increasing LBA order. Read batches take precedence over write batches by default, as
applications are more likely to block on read I/O. After a batch is processed, d ead l i ne
checks how long write operations have been starved of processor time and schedules the
next read or write batch as appropriate. The number of requests to handle per batch, the
number of read batches to issue per write batch, and the amount of time before requests
expire are all configurable; see Section 5.3.4, Tuning the deadline scheduler for details.
c f q
The default scheduler only for devices identified as SATA disks. The Completely Fair
Queueing scheduler, cfq , divides processes into three separate classes: real time, best
effort, and idle. Processes in the real time class are always performed before processes in
the best effort class, which are always performed before processes in the idle class. This
means that processes in the real time class can starve both best effort and idle processes of
processor time. Processes are assigned to the best effort class by default.
C fq uses historical data to anticipate whether an application will issue more I/O requests in
the near future. If more I/O is expected, cfq idles to wait for the new I/O, even if I/O from
other processes is waiting to be processed.
Because of this tendency to idle, the cfq scheduler should not be used in conjunction with
hardware that does not incur a large seek penalty unless it is tuned for this purpose. It
should also not be used in conjunction with other non-work-conserving schedulers, such
as a host-based hardware RAID controller, as stacking these schedulers tends to cause a
large amount of latency.
C fq behavior is highly configurable; see Section 5.3.5, Tuning the cfq scheduler for
details.
n o o p
The no o p I/O scheduler implements a simple FIFO (first-in first-out) scheduling algorithm.
Requests are merged at the generic block layer through a simple last-hit cache. This can be
the best scheduler for CPU-bound systems using fast storage.
For details on setting a different default I/O scheduler, or specifying a different scheduler for a
particular device, see Section 5.3, Configuration tools .
5.1.3. File syst ems

Read this section for details about supported file systems in Red Hat Enterprise Linux 7, their
recommended use cases, and the format and mount options available to file systems in general.
D etailed tuning recommendations for these file systems are available in Section 5.3.7, Configuring
file systems for performance .
34
5 .1 .3.1 . XFS
XFS is a robust and highly scalable 64-bit file system. It is the default file system in Red Hat
Enterprise Linux 7. XFS uses extent-based allocation, and features a number of allocation schemes,
including pre-allocation and delayed allocation, both of which reduce fragmentation and aid
performance. It also supports metadata journaling, which can facilitate crash recovery. XFS can be
defragmented and enlarged while mounted and active, and Red Hat Enterprise Linux 7 supports
several XFS-specific backup and restore utilities.
As of Red Hat Enterprise Linux 7.0 GA, XFS is supported to a maximum file system size of 500 TB,
and a maximum file offset of 8 EB (sparse files). For details about administering XFS, see the Red Hat
Enterprise Linux 7 Storage Administration Guide, available from
http://access.redhat.com/site/documentation/Red_Hat_Enterprise_Linux/. For assistance tuning XFS
for a specific purpose, see Section 5.3.7.1, Tuning XFS .
5 .1 .3.2 . Ext 4
Ext4 is a scalable extension of the ext3 file system. Its default behavior is optimal for most work loads.
However, it is supported only to a maximum file system size of 50 TB, and a maximum file size of
16 TB. For details about administering ext4, see the Red Hat Enterprise Linux 7 Storage Administration
Guide, available from http://access.redhat.com/site/documentation/Red_Hat_Enterprise_Linux/. For
assistance tuning ext4 for a specific purpose, see Section 5.3.7.2, Tuning ext4 .
5 .1 .3.3. Bt rfs (T e chno lo gy Pre vie w)

The current default file system for Red Hat Enterprise Linux 7 is XFS, however, a version of BTRFS is
also shipped as a technology preview. BTRFS is a relatively new copy-on-write file system. Some of
the unique functionality of BTRFS includes:
Ability to take snapshots of specific files, volumes or sub-volumes rather than the whole file
system.
Supports several versions of redundant array of inexpensive disks (RAID )
Back references map I/O errors to file system objects.
Transparent compression (all files on the partition are automatically compressed).
Checksums on data and meta-data.
Although BTRFS is considered a stable file system, being new it is under constant development, and
some functionality such as the repair tools are very basic compared to the maturity of other file
systems.
At the moment BTRFS is a good selection when advanced features (such as snapshots,
compression, and file data checksums) are required, but performance is relatively unimportant. If
these features are not required, then the risk of failure or poor performance over time make this a poor
choice for a file system. Another drawback, compared to other file systems, is the maximum supported
file system size of 50 TB.
Red Hat Enterprise Linux 7 provides btrfs as a Technology Preview. For details about Technology
Preview features, see https://access.redhat.com/site/support/offerings/techpreview/.
For details about administering BTRFs, see the Red Hat Enterprise Linux 7 Storage Administration Guide,
available from https://access.redhat.com/documentation/enUS/Red_Hat_Enterprise_Linux/7/html/Storage_Administration_Guide/ch-btrfs.html. For assistance
tuning BTRFS for a specific purpose, see section 5.3.7.3 Tuning BTRFS in http://documentationdevel.engineering.redhat.com/site/documentation/en-
35
US/Red_Hat_Enterprise_Linux/7/html/Performance_Tuning_Guide/sect-Red_Hat_Enterprise_LinuxPerformance_Tuning_Guide-Storage_and_File_Systems-Configuration_tools.html#sectRed_Hat_Enterprise_Linux-Performance_Tuning_Guide-Configuration_toolsConfiguring_file_systems_for_performance.
5 .1 .3.4 . GFS2
GFS2 is part of the High Availability Add-On, which provides clustered file system support to Red Hat
Enterprise Linux 7. GFS2 provides a consistent file system image across all servers in a cluster,
allowing servers to read from and write to a single shared file system.
GFS2 is supported to a maximum file system size of 250 TB.
For details about administering GFS2, see the Red Hat Enterprise Linux 7 Storage Administration Guide,
available from http://access.redhat.com/site/documentation/Red_Hat_Enterprise_Linux/. For
assistance tuning GFS2 for a specific purpose, see Section 5.3.7.4, Tuning GFS2 .
5.1.4 . Generic t uning considerat ions for file syst ems

This section covers tuning considerations common to all file systems. For tuning recommendations
specific to your file system, see Section 5.3.7, Configuring file systems for performance .
5 .1 .4 .1 . Co nside rat io ns at fo rm at t im e
Some file system configuration decisions cannot be changed after the device is formatted. This
section covers the options available to you for decisions that must be made before you format your
storage device.
S iz e
Create an appropriately-sized file system for your workload. Smaller file systems have
proportionally shorter backup times and require less time and memory for file system
checks. However, if your file system is too small, its performance will suffer from high
fragmentation.
B lo ck siz e
The block is the unit of work for the file system. The block size determines how much data
can be stored in a single block, and therefore the smallest amount of data that is written or
read at one time.
The default block size is appropriate for most use cases. However, your file system will
perform better and store data more efficiently if the block size (or the size of multiple blocks)
is the same as or slightly larger than amount of data that is typically read or written at one
time. A small file will still use an entire block. Files can be spread across multiple blocks, but
this can create additional runtime overhead. Additionally, some file systems are limited to a
certain number of blocks, which in turn limits the maximum size of the file system.
Block size is specified as part of the file system options when formatting a device with the
mkfs command. The parameter that specifies the block size varies with the file system; see
the mkfs man page for your file system for details. For example, to see the options available
when formatting an XFS file system, execute the following command.
$ man mkfs.xfs
G eo met ry
36
File system geometry is concerned with the distribution of data across a file system. If your
system uses striped storage, like RAID , you can improve performance by aligning data and
metadata with the underlying storage geometry when you format the device.
Many devices export recommended geometry, which is then set automatically when the
devices are formatted with a particular file system. If your device does not export these
recommendations, or you want to change the recommended settings, you must specify
geometry manually when you format the device with mkf s.
The parameters that specify file system geometry vary with the file system; see the mkfs man
page for your file system for details. For example, to see the options available when
formatting an ext4 file system, execute the following command.
$ man mkfs.ext4
E xt ern al jo u rn als
Journaling file systems document the changes that will be made during a write operation in
a journal file prior to the operation being executed. This reduces the likelihood that a
storage device will become corrupted in the event of a system crash or power failure, and
speeds up the recovery process.
Metadata-intensive workloads involve very frequent updates to the journal. A larger journal
uses more memory, but reduces the frequency of write operations. Additionally, you can
improve the seek time of a device with a metadata-intensive workload by placing its journal
on dedicated storage that is as fast as, or faster than, the primary storage.
Warning
Ensure that external journals are reliable. Losing an external journal device will
cause file system corruption.
External journals must be created at format time, with journal devices being specified at
mount time. For details, see the mkfs and mo unt man pages.
$ man mkfs
$ man mount
5 .1 .4 .2 . Co nside rat io ns at m o unt t im e

This section covers tuning decisions that apply to most file systems and can be specified as the
device is mounted.
B arriers
File system barriers ensure that file system metadata is correctly written and ordered on
persistent storage, and that data transmitted with fsync persists across a power outage.
On previous versions of Red Hat Enterprise Linux, enabling file system barriers could
significantly slow applications that relied heavily on fsync, or created and deleted many
small files.
In Red Hat Enterprise Linux 7, file system barrier performance has been improved such that
the performance effects of disabling file system barriers are negligible (less than 3% ).
37
For further information, see the Red Hat Enterprise Linux 7 Storage Administration Guide,
Access T ime
Every time a file is read, its metadata is updated with the time at which access occurred
(ati me). This involves additional write I/O. In most cases, this overhead is minimal, as by
default Red Hat Enterprise Linux 7 updates the ati me field only when the previous access
time was older than the times of last modification (mti me) or status change (cti me).
However, if updating this metadata is time consuming, and if accurate access time data is
not required, you can mount the file system with the no ati me mount option. This disables
updates to metadata when a file is read. It also enables no d i rati me behavior, which
disables updates to metadata when a directory is read.
R ead - ah ead
Read-ahead behavior speeds up file access by pre-fetching data that is likely to be needed
soon and loading it into the page cache, where it can be retrieved more quickly than if it
were on disk. The higher the read-ahead value, the further ahead the system pre-fetches
data.
Red Hat Enterprise Linux attempts to set an appropriate read-ahead value based on what it
detects about your file system. However, accurate detection is not always possible. For
example, if a storage array presents itself to the system as a single LUN, the system detects
the single LUN, and does not set the appropriate read-ahead value for an array.
Workloads that involve heavy streaming of sequential I/O often benefit from high readahead values. The storage-related tuned profiles provided with Red Hat Enterprise Linux 7
raise the read-ahead value, as does using LVM striping, but these adjustments are not
always sufficient for all workloads.
The parameters that define read-ahead behavior vary with the file system; see the mount
man page for details.
$ man mount
5 .1 .4 .3. Maint e nance

Regularly discarding blocks that are not in use by the file system is a recommended practice for both
solid-state disks and thinly-provisioned storage. There are two methods of discarding unused
blocks: batch discard and online discard.
B at ch d iscard
This type of discard is part of the f st rim command. It discards all unused blocks in a file
system that match criteria specified by the administrator.
Red Hat Enterprise Linux 7 supports batch discard on XFS and ext4 formatted devices that
support physical discard operations (that is, on HD D devices where the value of
/sys/bl o ck/devname/q ueue/d i scard _max_bytes is not zero, and SSD devices
where the value of /sys/bl o ck/sd a/q ueue/d i scard _g ranul ari ty is not 0 ).
O n lin e d iscard
38
This type of discard operation is configured at mount time with the d i scard option, and
runs in real time without user intervention. However, online discard only discards blocks
that are transitioning from used to free. Red Hat Enterprise Linux 7 supports online discard
on XFS and ext4 formatted devices.
Red Hat recommends batch discard except where online discard is required to maintain
performance, or where batch discard is not feasible for the system's workload.
P re- allo cat io n
Pre-allocation marks disk space as being allocated to a file without writing any data into
that space. This can be useful in limiting data fragmentation and poor read performance.
Red Hat Enterprise Linux 7 supports pre-allocating space on XFS, ext4, and GFS2 devices
at mount time; see the mo unt man page for the appropriate parameter for your file system.
Applications can also benefit from pre-allocating space by using the fal l o cate(2)
g l i bc call.

performance and diagnosing performance problems related to I/O and file systems and their
configuration. This section outlines the available tools and gives examples of how to use them to
monitor and diagnose I/O and file system related performance issues.
5.2.1. Monit oring syst em performance wit h vmst at

Vmstat reports on processes, memory, paging, block I/O, interrupts, and CPU activity across the
entire system. It can help administrators determine whether the I/O subsystem is responsible for any
performance issues.
The information most relevant to I/O performance is in the following columns:
s i
Swap in, or writes to swap space, in KB.
s o
Swap out, or reads from swap space, in KB.
b i
Block in, or block write operations, in KB.
b o
Block out, or block read operations, in KB.
w a
The portion of the queue that is waiting for I/O operations to complete.
Swap in and swap out are particularly useful when your swap space and your data are on the same
device, and as indicators of memory usage.
Additionally, the free, buff, and cache columns can help identify write-back frequency. A sudden drop
in cache values and an increase in free values indicates that write-back and page cache invalidation
has begun.
39
If analysis with vmst at shows that the I/O subsystem is responsible for reduced performance,
administrators can use io st at to determine the responsible I/O device.
vmst at is provided by the procps-ng package. For detailed information about using vmst at , see the
man page:
$ man vmstat
5.2.2. Monit oring I/O performance wit h iost at

Io st at is provided by the sysstat package. It reports on I/O device load in your system. If analysis
with vmst at shows that the I/O subsystem is responsible for reduced performance, you can use
io st at to determine the I/O device responsible.
You can focus the output of io st at reports on a specific device by using the parameters defined in
the io st at man page:
$ man iostat
5 .2 .2 .1 . De t aile d I/O analysis wit h blkt race

B lkt race provides detailed information about how time is spent in the I/O subsystem. The
companion utility b lkp arse reads the raw output from b lkt race and produces a human readable
summary of input and output operations recorded by b lkt race.
For more detailed information about this tool, see the man page:
$ man blktrace
$ man blkparse
5 .2 .2 .2 . Analyzing blkt race o ut put wit h bt t

B t t is provided as part of the blktrace package. It analyzes b lkt race output and displays the amount
of time that data spends in each area of the I/O stack, making it easier to spot bottlenecks in the I/O
subsystem.
For example, if b t t shows that the time between requests being sent to the block layer (Q 2Q ) is larger
than the total time that requests spent in the block layer (Q 2C ), the I/O subsystem may not be
responsible for performance issues. If the device takes a long time to service a request (D 2C ), the
device may be overloaded, or the workload sent to the device may be sub-optimal. If block I/O is
queued for a long time before being assigned a request (Q 2G ), it may indicate that the storage in use
is unable to serve the I/O load.
$ man btt
5 .2 .2 .3. Analyzing blkt race o ut put wit h se e kwat che r
40
The seekwat ch er tool can use b lkt race output to graph I/O over time. It focuses on the Logical
Block Address (LBA) of disk I/O, throughput in megabytes per second, the number of seeks per
second, and I/O operations per second. This can help you to identify when you are hitting the
operations-per-second limit of a device.
$ man seekwatcher
5.2.3. St orage monit oring wit h Syst emT ap

The Red Hat Enterprise Linux 7 SystemTap Beginner's Guide includes several sample scripts that are
useful for profiling and monitoring storage performance.
The following Syst emT ap example scripts relate to storage performance and may be useful in
diagnosing storage or file system performance problems. By default they are installed to the
/usr/share/d o c/systemtap-cl i ent/exampl es/i o directory.
d i skto p. stp
Checks the status of reading/writing disk every 5 seconds and outputs the top ten entries
during that period.
i o ti me. stp
Prints the amount of time spent on read and write operations, and the number of bytes read
and written.
tracei o . stp
Prints the top ten executables based on cumulative I/O traffic observed, every second.
tracei o 2. stp
Prints the executable name and process identifier as reads and writes to the specified
device occur.
i no d ewatch. stp
Prints the executable name and process identifier each time a read or write occurs to the
specified inode on the specified major/minor device.
i no d ewatch2. stp
Prints the executable name, process identifier, and attributes each time the attributes are
changed on the specified inode on the specified major/minor device.
The Red Hat Enterprise Linux 7 SystemTap Beginner's Guide is available from
5.3. Configurat ion t ools

storage and file systems. This section outlines the available tools and provides examples of how they
can be used to solve I/O and file system related performance problems in Red Hat Enterprise Linux 7.
5.3.1. Configuring t uning profiles for st orage performance
41
T u n ed and t u n ed - ad m provide a number of profiles designed to improve performance for specific

use cases. The following profiles are particularly useful for improving storage performance.
latency-performance
throughput-performance (the default)
To configure a profile on your system, run the following command, replacing name with the name of
the profile you want to use.
$ tuned-adm profile name
The tuned -ad m reco mmend command recommends an appropriate profile for your system. This
also sets the default profile for your system at install time, so can be used to return to the default
profile.
For further details about these profiles or additional configuration options, see Section A.6, tunedadm .
5.3.2. Set t ing t he default I/O scheduler

The default I/O scheduler is the scheduler that is used if no scheduler is specified in a device's mount
options.
To set the default I/O scheduler, specify the scheduler you want to use by appending the elevator
parameter to the kernel command line, either at boot time or by editing the /etc/g rub2. co nf file.
elevator=scheduler_name
5.3.3. Configuring t he I/O scheduler for a device

To set the scheduler or scheduler preference order for a particular storage device, edit the
/sys/bl o ck/devname/q ueue/sched ul er file, where devname is the name of the device you want
to configure.
# echo cfq > /sys/block/hda/queue/scheduler
5.3.4 . T uning t he deadline scheduler

When d ead l i ne is in use, queued I/O requests are sorted into a read or write batch and then
scheduled for execution in increasing LBA order. Read batches take precedence over write batches
by default, as applications are more likely to block on read I/O. After a batch is processed,
d ead l i ne checks how long write operations have been starved of processor time and schedules
the next read or write batch as appropriate.
The following parameters affect the behaviour of the d ead l i ne scheduler.
f if o _b at ch
The number of read or write operations to issue in a single batch. The default value is 16 . A
higher value can increase throughput, but will also increase latency.
f ro n t _merg es
If your workload will never generate front merges, this tunable can be set to 0 . However,
unless you have measured the overhead of this check, Red Hat recommends the default
42
unless you have measured the overhead of this check, Red Hat recommends the default
value of 1.
r ead _exp ire
The number of milliseconds in which a read request should be scheduled for service. The
default value is 50 0 (0.5 seconds).
w rit e_exp ire
The number of milliseconds in which a write request should be scheduled for service. The
default value is 50 0 0 (5 seconds).
w rit es_st arved
The number of read batches that can be processed before processing a write batch. The
higher this value is set, the greater the preference given to read batches.
5.3.5. T uning t he cfq scheduler

When cfq is in use, processes are placed into three classes: real time, best effort, and idle. All real
time processes are scheduled before any best effort processes, which are scheduled before any idle
processes. By default, processes are classed as best effort. You can manually adjust the class of a
process with the i o ni ce command.
You can further adjust the behavior of the cfq scheduler with the following parameters. These
parameters are set on a per-device basis by altering the specified files under the
/sys/bl o ck/devname/q ueue/i o sched directory.
b ack_seek_max
The maximum distance in kilobytes that cfq will perform a backward seek. The default
value is 16 KB. Backward seeks typically damage performance, so large values are not
recommended.
b ack_seek_p en alt y
The multiplier applied to backward seeks when the disk head is deciding whether to move
forward or backward. The default value is 2. If the disk head position is at 1024 KB, and
there are equidistant requests in the system (1008 KB and 1040 KB, for example), the
back_seek_penalty is applied to backward seek distances and the disk moves forward.
f if o _exp ire_asyn c
The length of time in milliseconds that an asynchronous (buffered write) request can remain
unserviced. After this amount of time expires, a single starved asynchronous request is
moved to the dispatch list. The default value is 250 milliseconds.
f if o _exp ire_syn c
The length of time in milliseconds that a synchronous (read or O _D IR EC T write) request
can remain unserviced. After this amount of time expires, a single starved synchronous
request is moved to the dispatch list. The default value is 125 milliseconds.
g ro u p _id le
This parameter is set to 0 (disabled) by default. When set to 1 (enabled), the cfq scheduler
idles on the last process that is issuing I/O in a control group. This is useful when using
proportional weight I/O control groups and when slice_idle is set to 0 (on fast storage).
g ro u p _iso lat io n
43
This parameter is set to 0 (disabled) by default. When set to 1 (enabled), it provides

stronger isolation between groups, but reduces throughput, as fairness is applied to both
random and sequential workloads. When group_isolation is disabled (set to 0 ),
fairness is provided to sequential workloads only. For more information, see the installed
documentation in /usr/share/d o c/kernel -d o cversi o n/D o cumentati o n/cg ro ups/bl ki o -co ntro l l er. txt.
l o w_lat en cy
This parameter is set to 1 (enabled) by default. When enabled, cfq favors fairness over
throughput by providing a maximum wait time of 30 0 ms for each process issuing I/O on a
device. When this parameter is set to 0 (disabled), target latency is ignored and each
process receives a full time slice.
q u an t u m
This parameter defines the number of I/O requests that cfq sends to one device at one time,
essentially limiting queue depth. The default value is 8 requests. The device being used
may support greater queue depth, but increasing the value of quantum will also increase
latency, especially for large sequential write work loads.
s lice_asyn c
This parameter defines the length of the time slice (in milliseconds) allotted to each process
issuing asynchronous I/O requests. The default value is 4 0 milliseconds.
s lice_id le
This parameter specifies the length of time in milliseconds that cfq idles while waiting for
further requests. The default value is 0 (no idling at the queue or service tree level). The
default value is ideal for throughput on external RAID storage, but can degrade throughput
on internal non-RAID storage as it increases the overall number of seek operations.
s lice_syn c
This parameter defines the length of the time slice (in milliseconds) allotted to each process
issuing synchronous I/O requests. The default value is 10 0 ms.
5 .3.5 .1 . T uning cfq fo r fast st o rage

The cfq scheduler is not recommended for hardware that does not suffer a large seek penalty, such
as fast external storage arrays or solid state disks. If your use case requires cfq to be used on this
storage, you will need to edit the following configuration files:
Set /sys/bl o ck/devname/q ueue/i o ni ce/sl i ce_i d l e to 0
Set /sys/bl o ck/devname/q ueue/i o ni ce/q uantum to 6 4
Set /sys/bl o ck/devname/q ueue/i o ni ce/g ro up_i d l e to 1
5.3.6. T uning t he noop scheduler

The no o p I/O scheduler is primarily useful for CPU-bound systems using fast storage. Requests are
merged at the block layer, so no o p behavior is modified by editing block layer parameters in the files
under the /sys/bl o ck/sd X/q ueue/ directory.
a d d _ran d o m
Some I/O events contribute to the entropy pool for /d ev/rand o m. This parameter can be
set to 0 if the overhead of these contributions becomes measurable.
44
set to 0 if the overhead of these contributions becomes measurable.

max_sect o rs_kb
Specifies the maximum size of an I/O request in kilobytes. The default value is 512 KB. The
minimum value for this parameter is determined by the logical block size of the storage
device. The maximum value for this parameter is determined by the value of
max_hw_sectors_kb.
Some solid state disks perform poorly when I/O requests are larger than the internal erase
block size. In these cases, Red Hat recommends reducing max_hw_sectors_kb to the
internal erase block size.
n o merg es
Most workloads benefit from request merging. However, disabling merges can be useful for
debugging purposes. Set this parameter to 0 to disable merging. It is enabled (set to 1) by
default.
n r_req u est s
Specifies the maximum number of read and write requests that can be queued at one time.
The default value is 128; that is, 128 read requests and 128 write requests can be queued
before the next process to request a read or write is put to sleep.
For latency-senstitive applications, lower the value of this parameter and limit the command
queue depth on the storage so that write-back I/O cannot fill the device queue with write
requests. When the device queue fills, other processes attempting to perform I/O operations
are put to sleep until queue space becomes available. Requests are then allocated in a
round-robin fashion, preventing one process from continuously consuming all spots in the
queue.
o p t imal_io _siz e
Some storage devices report an optimal I/O size through this parameter. If this value is
reported, Red Hat recommends that applications issue I/O aligned to and in multiples of the
optimal I/O size wherever possible.
r ead _ah ead _kb
D efines the number of kilobytes that the operating system will read ahead during a
sequential read operation in order to store information likely to be needed soon in the page
cache. D evice mappers often benefit from a high read_ahead_kb value; 128 KB for each
device to be mapped is a good starting point.
r o t at io n al
Some solid state disks do not correctly advertise their solid state status, and are mounted
as traditional rotational disks. If your solid state device does does not set this to 0
automatically, set it manually to disable unnecessary seek-reducing logic in the scheduler.
r q _af f in it y
By default, I/O completions can be processed on a different processor than the processor
that issued the I/O request. Set rq_affinity to 1 to disable this ability and perform
completions only on the processor that issued the I/O request. This can improve the
effectiveness of processor data caching.
5.3.7. Configuring file syst ems for performance
45
This section covers the tuning parameters specific to each file system supported in Red Hat
Enterprise Linux 7. Parameters are divided according to whether their values should be configured
when you format the storage device, or when you mount the formatted device.
Where loss in performance is caused by file fragmentation or resource contention, performance can
generally be improved by reconfiguring the file system. However, in some cases the application may
need to be altered. In this case, Red Hat recommends contacting Customer Support for assistance.
5 .3.7 .1 . T uning XFS

This section covers some of the tuning parameters available to XFS file systems at format and at
mount time.
The default formatting and mount settings for XFS are suitable for most workloads. Red Hat
recommends changing them only if specific configuration changes are expected to benefit your
workload.
5.3.7.1.1. Fo rmat t in g o p t io n s
For further details about any of these formatting options, see the man page:
$ man mkfs.xfs
D irect o ry b lo ck siz e
The directory block size affects the amount of directory information that can be retrieved or
modified per I/O operation. The minimum value for directory block size is the file system
block size (4 KB by default). The maximum value for directory block size is 6 4 KB.
At a given directory block size, a larger directory requires more I/O than a smaller directory.
A system with a larger directory block size also consumes more processing power per I/O
operation than a system with a smaller directory block size. It is therefore recommended to
have as small a directory and directory block size as possible for your workload.
Red Hat recommends the directory block sizes listed in Table 5.1, Recommended maximum
directory entries for directory block sizes for file systems with no more than the listed
number of entries for write-heavy and read-heavy workloads.
T ab le 5.1. R eco mmen d ed maximu m d irect o ry en t ries f o r d irect o ry b lo ck siz es
D irect o ry b lo ck siz e
Max. en t ries ( read h eavy)
Max. en t ries ( writ eh eavy)
4 KB
16 KB
64 KB
100000200000
1000001000000
>1000000
10000002000000
100000010000000
>10000000
For detailed information about the effect of directory block size on read and write workloads
in file systems of different sizes, see the XFS documentation.
To configure directory block size, use the mkfs. xfs -l option. See the mkfs. xfs man
page for details.
Allo cat io n g ro u p s
An allocation group is an independent structure that indexes free space and allocated
inodes across a section of the file system. Each allocation group can be modified
independently, allowing XFS to perform allocation and deallocation operations
46
concurrently as long as concurrent operations affect different allocation groups. The

number of concurrent operations that can be performed in the file system is therefore equal
to the number of allocation groups. However, since the ability to perform concurrent
operations is also limited by the number of processors able to perform the operations,
Red Hat recommends that the number of allocation groups be greater than or equal to the
number of processors in the system.
A single directory cannot be modified by multiple allocation groups simultaneously.
Therefore, Red Hat recommends that applications that create and remove large numbers of
files do not store all files in a single directory.
To configure allocation groups, use the mkfs. xfs -d option. See the mkfs. xfs man
page for details.
G ro wt h co n st rain t s
If you may need to increase the size of your file system after formatting time (either by
adding more hardware or through thin-provisioning), you must carefully consider initial file
layout, as allocation group size cannot be changed after formatting is complete.
Allocation groups must be sized according to the eventual capacity of the file system, not
the initial capacity. The number of allocation groups in the fully-grown file system should
not exceed several hundred, unless allocation groups are at their maximum size (1 TB).
Therefore for most file systems, the recommended maximum growth to allow for a file system
is ten times the initial size.
Additional care must be taken when growing a file system on a RAID array, as the device
size must be aligned to an exact multiple of the allocation group size so that new allocation
group headers are correctly aligned on the newly added storage. The new storage must
also have the same geometry as the existing storage, since geometry cannot be changed
after formatting time, and therefore cannot be optimized for storage of a different geometry
on the same block device.
I n o d e siz e an d in lin e at t rib u t es
If the inode has sufficient space available, XFS can write attribute names and values
directly into the inode. These inline attributes can be retrieved and modified up to an order
of magnitude faster than retrieving separate attribute blocks, as additional I/O is not
required.
The default inode size is 256 bytes. Only around 100 bytes of this is available for attribute
storage, depending on the number of data extent pointers stored in the inode. Increasing
inode size when you format the file system can increase the amount of space available for
storing attributes.
Both attribute names and attribute values are limited to a maximum size of 254 bytes. If
either name or value exceeds 254 bytes in length, the attribute is pushed to a separate
attribute block instead of being stored inline.
To configure inode parameters, use the mkfs. xfs -i option. See the mkfs. xfs man
page for details.
R AID
If software RAID is in use, mkfs. xfs automatically configures the underlying hardware
with an appropriate stripe unit and width. However, stripe unit and width may need to be
manually configured if hardware RAID is in use, as not all hardware RAID devices export
this information. To configure stripe unit and width, use the mkfs. xfs -d option. See the
mkfs. xfs man page for details.
47
L o g siz e
Pending changes are aggregated in memory until a synchronization event is triggered, at
which point they are written to the log. The size of the log determines the number of
concurrent modifications that can be in-progress at one time. It also determines the
maximum amount of change that can be aggregated in memory, and therefore how often
logged data is written to disk. A smaller log forces data to be written back to disk more
frequently than a larger log. However, a larger log uses more memory to record pending
modifications, so a system with limited memory will not benefit from a larger log.
Logs perform better when they are aligned to the underlying stripe unit; that is, they start
and end at stripe unit boundaries. To align logs to the stripe unit, use the mkfs. xfs -d
option. See the mkfs. xfs man page for details.
To configure the log size, use the following mkfs. xfs option, replacing logsize with the
size of the log:
# mkfs.xfs -l size=logsize
For further details, see the mkfs. xfs man page:
$ man mkfs.xfs
L o g st rip e u n it
Log writes on storage devices that use RAID 5 or RAID 6 layouts may perform better when
they start and end at stripe unit boundaries (are aligned to the underlying stripe unit).
mkfs. xfs attempts to set an appropriate log stripe unit automatically, but this depends on
the RAID device exporting this information.
Setting a large log stripe unit can harm performance if your workload triggers
synchronization events very frequently, because smaller writes need to be padded to the
size of the log stripe unit, which can increase latency. If your workload is bound by log write
latency, Red Hat recommends setting the log stripe unit to 1 block so that unaligned log
writes are triggered as possible.
The maximum supported log stripe unit is the size of the maximum log buffer size (256 KB).
It is therefore possible that the underlying storage may have a larger stripe unit than can be
configured on the log. In this case, mkfs. xfs issues a warning and sets a log stripe unit of
32 KB.
To configure the log stripe unit, use one of the following options, where N is the number of
blocks to use as the stripe unit, and size is the size of the stripe unit in KB.
mkfs.xfs -l sunit=Nb
mkfs.xfs -l su=size
For further details, see the mkfs. xfs man page:
$ man mkfs.xfs
5.3.7.1.2. Mo u n t o p t io n s
I n o d e allo cat io n
48
compress=zlib - Better compression ratio. It's the default and safe for older kernels.
(default).
compress=lzo - Faster compression but does not compress as much as zlib.
compress=no - D isables compression (starting with kernel 3.6).
compress-force=method - Enable compression even for files that don't compress well,
like videos and dd images of disks. The options are compress-force=zlib and compressforce=lzo.
Highly recommended for file systems greater than 1 TB in size. The inode64 parameter
configures XFS to allocate inodes and data across the entire file system. This ensures that
inodes are not allocated largely at the beginning of the file system, and data is not largely
allocated at the end of the file system, improving performance on large file systems.
L o g b u f f er siz e an d n u mb er
The larger the log buffer, the fewer I/O operations it takes to write all changes to the log. A
larger log buffer can improve performance on systems with I/O-intensive workloads that do
not have a non-volatile write cache.
The log buffer size is configured with the logbsize mount option, and defines the
maximum amount of information that can be stored in the log buffer; if a log stripe unit is not
set, buffer writes can be shorter than the maximum, and therefore there is no need to reduce
the log buffer size for synchronization-heavy workloads. The default size of the log buffer is
32 KB. The maximum size is 256 KB and other supported sizes are 64 KB, 128 KB or power
of 2 multiples of the log stripe unit between 32 KB and 256 KB.
The number of log buffers is defined by the logbufs mount option. The default value is 8
log buffers (the maximum), but as few as two log buffers can be configured. It is usually not
necessary to reduce the number of log buffers, except on memory-bound systems that
cannot afford to allocate memory to additional log buffers. Reducing the number of log
buffers tends to reduce log performance, especially on workloads sensitive to log I/O
latency.
D elay ch an g e lo g g in g
XFS has the option to aggregate changes in memory before writing them to the log. The
delaylog parameter allows frequently modified metadata to be written to the log
periodically instead of every time it changes. This option increases the potential number of
operations lost in a crash and increases the amount of memory used to track metadata.
However, it can also increase metadata modification speed and scalability by an order of
magnitude, and does not reduce data or metadata integrity when fsync, fd atasync, or
sync are used to ensure data and metadata is written to disk.
5 .3.7 .2 . T uning e xt 4
This section covers some of the tuning parameters available to ext4 file systems at format and at
mount time.
5.3.7.2.1. Fo rmat t in g o p t io n s
I n o d e t ab le in it ializ at io n
Initializing all inodes in the file system can take a very long time on very large file systems.
By default, the initialization process is deferred (lazy inode table initialization is enabled).
However, if your system does not have an ext4 driver, lazy inode table initialization is
49
disabled by default. It can be enabled by setting l azy_i tabl e_i ni t to 1). In this case,
kernel processes continue to initialize the file system after it is mounted.
This section describes only some of the options available at format time. For further formatting
parameters, see the mkfs. ext4 man page:
$ man mkfs.ext4
5.3.7.2.2. Mo u n t o p t io n s
I n o d e t ab le in it ializ at io n rat e
When lazy inode table initialization is enabled, you can control the rate at which
initialization occurs by specifying a value for the init_itable parameter. The amount of
time spent performing background initialization is approximately equal to 1 divided by the
value of this parameter. The default value is 10 .
Au t o mat ic f ile syn ch ro n iz at io n
Some applications do not correctly perform an fsync after renaming an existing file, or
after truncating and rewriting. By default, ext4 automatically synchronizes files after each of
these operations. However, this can be time consuming.
If this level of synchronization is not required, you can disable this behavior by specifying
the no auto _d a_al l o c option at mount time. If no auto _d a_al l o c is set, applications
must explicitly use fsync to ensure data persistence.
J o u rn al I/O p rio rit y
By default, journal I/O has a priority of 3, which is slightly higher than the priority of normal
I/O. You can control the priority of journal I/O with the journal_ioprio parameter at
mount time. Valid values for journal_ioprio range from 0 to 7, with 0 being the highest
priority I/O.
This section describes only some of the options available at mount time. For further mount options,
see the mo unt man page:
$ man mount
5 .3.7 .3. T uning BT RFS

As of Red Hat Enterprise Linux 7.0, BTRFS is provided as a Technology Preview. Tuning should
always be done to optimize the system based on it's current workload. This section deals only with
optimizing a BTRFS, for creation and mounting options see the Red Hat Enterprise Linux 7 Storage
Administration Guide, available from https://access.redhat.com/documentation/enUS/Red_Hat_Enterprise_Linux/7/html/Storage_Administration_Guide/ch-btrfs.html
5.3.7.3.1. D at a C o mp ressio n
The default compression algorithm is zlib, but depending on a typical workload there many be a
reason to change it. For example, if you have a single thread with heavy file I/O, you may want to use
the lzo algorithm. Options at mount time are:
compress=zlib - Better compression ratio. It's the default and safe for older kernels.(default).
compress=lzo - Faster compression but does not compress as much as zlib.
50
compress=no - D isables compression (starting with kernel 3.6).

compress-force=method - Enable compression even for files that don't compress well, like videos
and images of disks. The options are compress-force=zlib and compress-force=lzo.
Only files created or changed after the mount option is added will be compressed. To compress
existing files use:
$ btrfs filesystem defragment -calg
alg is either zl i b or l zo
To re-compress the file using l zo use:
$ btrfs filesystem defragment -r -v -clzo /
5 .3.7 .4 . T uning GFS2

This section covers some of the tuning parameters available to GFS2 file systems at format and at
mount time.
D irect o ry sp acin g
All directories created in the top-level directory of the GFS2 mount point are automatically
spaced to reduce fragmentation and increase write speed in those directories. To space
another directory like a top-level directory, mark that directory with the T attribute, as shown,
replacing dirname with the path to the directory you wish to space:
# chattr +T dirname
chattr is provided as part of the e2fsprogs package.
R ed u ce co n t en t io n
GFS2 uses a global locking mechanism that can require communication between the
nodes of a cluster. Contention for files and directories between multiple nodes lowers
performance. You can minimize the risk of cross-cache invalidation by minimizing the areas
of the file system that are shared between multiple nodes.
51
Chapter 6. Networking
The networking subsystem is comprised of a number of different parts with sensitive connections.
Red Hat Enterprise Linux 7 networking is therefore designed to provide optimal performance for most
workloads, and to optimize its performance automatically. As such, it is not usually necessary to
manually tune network performance. This chapter discusses further optimizations that can be made
to functional networking systems.
Network performance problems are sometimes the result of hardware malfunction or faulty
infrastructure. Resolving these issues is beyond the scope of this document.

To make good tuning decisions, you need a thorough understanding of packet reception in Red Hat
Enterprise Linux. This section explains how network packets are received and processed, and where
potential bottlenecks can occur.
A packet sent to a Red Hat Enterprise Linux system is received by the network interface card (NIC)
and placed in either an internal hardware buffer or a ring buffer. The NIC then sends a hardware
interrupt request, prompting the creation of a software interrupt operation to handle the interrupt
request.
As part of the software interrupt operation, the packet is transferred from the buffer to the network
stack. D epending on the packet and your network configuration, the packet is then forwarded,
discarded, or passed to a socket receive queue for an application and then removed from the
network stack. This process continues until either there are no packets left in the NIC hardware buffer,
or a certain number of packets (specified in /pro c/sys/net/co re/d ev_wei g ht) are transferred.
6.1.1. Before you t une

Network performance problems are most often the result of hardware malfunction or faulty
infrastructure. Red Hat highly recommends verifying that your hardware and infrastructure are
working as expected before beginning to tune the network stack.
6.1.2. Bot t lenecks in packet recept ion

While the network stack is largely self-optimizing, there are a number of points during network packet
processing that can become bottlenecks and reduce performance.
T h e N IC h ard ware b u f f er o r rin g b u f f er
The hardware buffer might be a bottleneck if a large number of packets are being dropped.
For information about monitoring your system for dropped packets, see Section 6.2.4,
ethtool .
T h e h ard ware o r so f t ware in t erru p t q u eu es
Interrupts can increase latency and processor contention. For information on how
interrupts are handled by the processor, see Section 3.1.3, Interrupt Request (IRQ)
Handling . For information on how to monitor interrupt handling in your system, see
Section 3.2.3, /proc/interrupts . For configuration options that affect interrupt handling,
see Section 3.3.7, Setting interrupt affinity .
T h e so cket receive q u eu e f o r t h e ap p licat io n
A bottleneck in an application's receive queue is indicated by a large number of packets
52
Chapt er 6 . Net working
that are not copied to the requesting application, or by an increase in UD P input errors
(InErro rs) in /pro c/net/snmp. For information about monitoring your system for these
errors, see Section 6.2.1, ss and Section 6.2.5, /proc/net/snmp .

performance and diagnosing performance problems related to the networking subsystem. This
section outlines the available tools and gives examples of how to use them to monitor and diagnose
network related performance issues.
6.2.1. ss
to assess device performance over time. By default, ss lists open non-listening TCP sockets that
have established connections, but a number of useful options are provided to help administrators
filter out statistics about specific sockets.
Red Hat recommends ss over n et st at in Red Hat Enterprise Linux 7.
$ man ss
6.2.2. ip
The ip utility lets administrators manage and monitor routes, devices, routing policies, and tunnels.
The i p mo ni to r command can continuously monitor the state of devices, addresses, and routes.
ip is provided by the iproute package. For details about using ip, see the man page:
$ man ip
6.2.3. dropwat ch
D ro p wat ch is an interactive tool that monitors and records packets that are dropped by the kernel.
For further information, see the d ro p wat ch man page:
$ man dropwatch
6.2.4 . et ht ool
The et h t o o l utility allows administrators to view and edit network interface card settings. It is useful
for observing the statistics of certain devices, such as the number of packets dropped by that device.
You can view the status of a specified device's counters with ethto o l -S and the name of the
device you want to monitor.
$ ethtool -S devname
For further information, see the man page:
53
$ man ethtool
6.2.5. /proc/net /snmp

The /pro c/net/snmp file displays data that is used by snmp agents for IP, ICMP, TCP and UD P
monitoring and management. Examining this file on a regular basis can help administrators identify
unusual values and thereby identify potential performance problems. For example, an increase in
UD P input errors (InErro rs) in /pro c/net/snmp can indicate a bottleneck in a socket receive
queue.
6.2.6. Net work monit oring wit h Syst emT ap

The Red Hat Enterprise Linux 7 SystemTap Beginner's Guide includes several sample scripts that are
useful for profiling and monitoring network performance.
The following Syst emT ap example scripts relate to networking and may be useful in diagnosing
network performance problems. By default they are installed to the /usr/share/d o c/systemtapcl i ent/exampl es/netwo rk directory.
n etto p. stp
Every 5 seconds, prints a list of processes (process identifier and command) with the
number of packets sent and received and the amount of data sent and received by the
process during that interval.
s o cket-trace. stp
Instruments each of the functions in the Linux kernel's net/so cket. c file, and prints trace
data.
d ro pwatch. stp
Every 5 seconds, prints the number of socket buffers freed at locations in the kernel. Use the
--al l -mo d ul es option to see symbolic names.
The l atencytap. stp script records the effect that different types of latency have on one or more
processes. It prints a list of latency types every 30 seconds, sorted in descending order by the total
time the process or processes spent waiting. This can be useful for identifying the cause of both
storage and network latency. Red Hat recommends using the --al l -mo d ul es option with this
script to better enable the mapping of latency events. By default, this script is installed to the
/usr/share/d o c/systemtap-cl i ent-version/exampl es/pro fi l i ng directory.
For further information, see the Red Hat Enterprise Linux 7 SystemTap Beginner's Guide, available from
6.3. Configurat ion t ools

system. This section outlines the available tools and provides examples of how they can be used to
solve network related performance problems in Red Hat Enterprise Linux 7.
However, it is important to keep in mind that network performance problems are sometimes the result
of hardware malfunction or faulty infrastructure. Red Hat highly recommends verifying that your
hardware and infrastructure are working as expected before using these tools to tune the network
stack.
54
Further, some network performance problems are better resolved by altering the application than by
reconfiguring your network subsystem. It is generally a good idea to configure your application to
perform frequent posix calls, even if this means queuing data in the application space, as this allows
data to be stored flexibly and swapped in or out of memory as required.
6.3.1. T uned-adm profiles for net work performance

t u n ed - ad m provides a number of different profiles to improve performance in a number of specific
use cases. The following profiles can be useful for improving networking performance.
latency-performance
network-latency
network-throughput
For more information about these profiles, see Section A.6, tuned-adm .
6.3.2. Configuring t he hardware buffer

If a large number of packets are being dropped by the hardware buffer, there are a number of
potential solutions.
S lo w t h e in p u t t raf f ic
Filter incoming traffic, reduce the number of joined multicast groups, or reduce the amount
of broadcast traffic to decrease the rate at which the queue fills. For details of how to filter
incoming traffic, see the Red Hat Enterprise Linux 7 Security Guide. For details about multicast
groups, see the Red Hat Enterprise Linux 7 Clustering documentation. For details about
broadcast traffic, see the Red Hat Enterprise Linux 7 System Administrator's Reference Guide, or
documentation related to the device you want to configure. All Red Hat Enterprise Linux 7
documentation is available from
R esiz e t h e h ard ware b u f f er q u eu e
Reduce the number of packets being dropped by increasing the size of the queue so that
the it does not overflow as easily. You can modify the rx/tx parameters of the network device
with the ethtool command:
# ethtool --set-ring devname value
C h an g e t h e d rain rat e o f t h e q u eu e
D evice weight refers to the number of packets a device can receive at one time (in a single
scheduled processor access). You can increase the rate at which a queue is drained by
increasing its device weight, which is controlled by the dev_weight parameter. This
parameter can be temporarily altered by changing the contents of the
/pro c/sys/net/co re/d ev_wei g ht file, or permanently altered with sysct l, which is
provided by the procps-ng package.
Altering the drain rate of a queue is usually the simplest way to mitigate poor network performance.
However, increasing the number of packets that a device can receive at one time uses additional
processor time, during which no other processes can be scheduled, so this can cause other
performance problems.
6.3.3. Configuring int errupt queues
55
If analysis reveals high latency, your system may benefit from poll-based rather than interrupt-based
packet receipt.
6 .3.3.1 . Co nfiguring busy po lling

Busy polling helps reduce latency in the network receive path by allowing socket layer code to poll
the receive queue of a network device, and disabling network interrupts. This removes delays caused
by the interrupt and the resultant context switch. However, it also increases CPU utilization. Busy
polling also prevents the CPU from sleeping, which can incur additional power consumption.
Busy polling is disabled by default. To enable busy polling on specific sockets, do the following.
Set sysctl.net.core.busy_poll to a value other than 0 . This parameter controls the number
of microseconds to wait for packets on the device queue for socket poll and selects. Red Hat
recommends a value of 50 .
Add the SO _BUSY _P O LL socket option to the socket.
To enable busy polling globally, you must also set sysctl.net.core.busy_read to a value other
than 0 . This parameter controls the number of microseconds to wait for packets on the device queue
for socket reads. It also sets the default value of the SO _BUSY _P O LL option. Red Hat recommends a
value of 50 for a small number of sockets, and a value of 10 0 for large numbers of sockets. For
extremely large numbers of sockets (more than several hundred), use epo l l instead.
Busy polling behavior is supported by the following drivers. These drivers are also supported on
Red Hat Enterprise Linux 7.
bnx2x
be2net
ixgbe
mlx4
myri10ge
As of Red Hat Enterprise Linux 7.1, you can also run the following command to check whether a
specific device supports busy polling.
# ethtool -k device | grep "busy-poll"
If this returns busy-po l l : o n [fi xed ], busy polling is available on the device.
6.3.4 . Configuring socket receive queues

If analysis suggests that packets are being dropped because the drain rate of a socket queue is too
slow, there are several ways to alleviate the performance issues that result.
D ecrease t h e sp eed o f in co min g t raf f ic
D ecrease the rate at which the queue fills by filtering or dropping packets before they reach
the queue, or by lowering the weight of the device.
I n crease t h e d ep t h o f t h e ap p licat io n ' s so cket q u eu e
56
If a socket queue that receives a limited amount of traffic in bursts, increasing the depth of
the socket queue to match the size of the bursts of traffic may prevent packets from being
dropped.
6 .3.4 .1 . De cre ase t he spe e d o f inco m ing t raffic

Filter incoming traffic or lower the network interface card's device weight to slow incoming traffic. For
details of how to filter incoming traffic, see the Red Hat Enterprise Linux 7 Security Guide, available from
D evice weight refers to the number of packets a device can receive at one time (in a single scheduled
processor access). D evice weight is controlled by the dev_weight parameter. This parameter can
be temporarily altered by changing the contents of the /pro c/sys/net/co re/d ev_wei g ht file, or
permanently altered with sysct l, which is provided by the procps-ng package.
6 .3.4 .2 . Incre asing que ue de pt h

Increasing the depth of an application socket queue is typically the easiest way to improve the drain
rate of a socket queue, but it is unlikely to be a long-term solution.
To increase the depth of a queue, increase the size of the socket receive buffer by making either of the
following changes:
I n crease t h e valu e o f /p ro c/sys/n et /co re/rmem_d ef au lt
This parameter controls the default size of the receive buffer used by sockets. This value
must be smaller than that of /pro c/sys/net/co re/rmem_max.
U se set so cko p t t o co n f ig u re a larg er SO _R C VB U F valu e
This parameter controls the maximum size in bytes of a socket's receive buffer. Use the
g etso cko pt system call to determine the current value of the buffer. For further information
about this parameter, see the man page:
$ man 7 socket
6.3.5. Configuring Receive-Side Scaling (RSS)

Receive-Side Scaling (RSS), also known as multi-queue receive, distributes network receive
processing across several hardware-based receive queues, allowing inbound network traffic to be
processed by multiple CPUs. RSS can be used to relieve bottlenecks in receive interrupt processing
caused by overloading a single CPU, and to reduce network latency.
To determine whether your network interface card supports RSS, check whether multiple interrupt
request queues are associated with the interface in /pro c/i nterrupts. For example, if you are
interested in the p1p1 interface:
# egrep 'CPU|p1p1' /proc/interrupts
CPU0
CPU1
CPU2
CPU3
CPU4
89:
40187
0
0
0
p1p1-0
90:
0
790
0
0
p1p1-1
91:
0
0
959
0
p1p1-2
92:
0
0
0
3310
CPU5
0
IR-PCI-MSI-edge
IR-PCI-MSI-edge
IR-PCI-MSI-edge
IR-PCI-MSI-edge
57
p1p1-3
93:
p1p1-4
94:
p1p1-5
622
IR-PCI-MSI-edge
2475
IR-PCI-MSI-edge
The preceding output shows that the NIC driver created 6 receive queues for the p1p1 interface
(p1p1-0 through p1p1-5). It also shows how many interrupts were processed by each queue, and
which CPU serviced the interrupt. In this case, there are 6 queues because by default, this particular
NIC driver creates one queue per CPU, and this system has 6 CPUs. This is a fairly common pattern
amongst NIC drivers.
Alternatively, you can check the output of l s -1
/sys/d evi ces/*/*/device_pci_address/msi _i rq s after the network driver is loaded. For
example, if you are interested in a device with a PCI address of 0 0 0 0 : 0 1: 0 0 . 0 , you can list the
interrupt request queues of that device with the following command:
# ls -1 /sys/devices/*/*/0000:01:00.0/msi_irqs
101
102
103
104
105
106
107
108
109
RSS is enabled by default. The number of queues (or the CPUs that should process network activity)
for RSS are configured in the appropriate network device driver. For the bnx2x driver, it is configured
in num_queues. For the sfc driver, it is configured in the rss_cpus parameter. Regardless, it is
typically configured in /sys/cl ass/net/device/q ueues/rx-queue/, where device is the name
of the network device (such as eth1) and rx-queue is the name of the appropriate receive queue.
When configuring RSS, Red Hat recommends limiting the number of queues to one per physical CPU
core. Hyper-threads are often represented as separate cores in analysis tools, but configuring
queues for all cores including logical cores such as hyper-threads has not proven beneficial to
network performance.
When enabled, RSS distributes network processing equally between available CPUs based on the
amount of processing each CPU has queued. However, you can use the ethto o l --show-rxfhindir and --set-rxfh-indir parameters to modify how network activity is distributed, and weight
certain types of network activity as more important than others.
The i rq bal ance daemon can be used in conjunction with RSS to reduce the likelihood of crossnode memory transfers and cache line bouncing. This lowers the latency of processing network
packets.
6.3.6. Configuring Receive Packet St eering (RPS)

Receive Packet Steering (RPS) is similar to RSS in that it is used to direct packets to specific CPUs
for processing. However, RPS is implemented at the software level, and helps to prevent the hardware
queue of a single network interface card from becoming a bottleneck in network traffic.
RPS has several advantages over hardware-based RSS:
RPS can be used with any network interface card.
58
It is easy to add software filters to RPS to deal with new protocols.

RPS does not increase the hardware interrupt rate of the network device. However, it does
introduce inter-processor interrupts.
RPS is configured per network device and receive queue, in the
/sys/cl ass/net/device/q ueues/rx-queue/rps_cpus file, where device is the name of the
network device (such as eth0 ) and rx-queue is the name of the appropriate receive queue (such as
rx-0 ).
The default value of the rps_cpus file is 0 . This disables RPS, so the CPU that handles the network
interrupt also processes the packet.
To enable RPS, configure the appropriate rps_cpus file with the CPUs that should process packets
from the specified network device and receive queue.
The rps_cpus files use comma-delimited CPU bitmaps. Therefore, to allow a CPU to handle
interrupts for the receive queue on an interface, set the value of their positions in the bitmap to 1. For
example, to handle interrupts with CPUs 0, 1, 2, and 3, set the value of rps_cpus to 0 0 0 0 1111
(1+2+4+8), or f (the hexadecimal value for 15).
For network devices with single transmit queues, best performance can be achieved by configuring
RPS to use CPUs in the same memory domain. On non-NUMA systems, this means that all available
CPUs can be used. If the network interrupt rate is extremely high, excluding the CPU that handles
network interrupts may also improve performance.
For network devices with multiple queues, there is typically no benefit to configuring both RPS and
RSS, as RSS is configured to map a CPU to each receive queue by default. However, RPS may still
be beneficial if there are fewer hardware queues than CPUs, and RPS is configured to use CPUs in
the same memory domain.
6.3.7. Configuring Receive Flow St eering (RFS)

Receive Flow Steering (RFS) extends RPS behavior to increase the CPU cache hit rate and thereby
reduce network latency. Where RPS forwards packets based solely on queue length, RFS uses the
RPS backend to calculate the most appropriate CPU, then forwards packets based on the location of
the application consuming the packet. This increases CPU cache efficiency.
RFS is disabled by default. To enable RFS, you must edit two files:
/ pro c/sys/net/co re/rps_so ck_fl o w_entri es
Set the value of this file to the maximum expected number of concurrently active
connections. We recommend a value of 3276 8 for moderate server loads. All values entered
are rounded up to the nearest power of 2 in practice.
/ sys/cl ass/net/device/q ueues/rx-queue/rps_fl o w_cnt
Replace device with the name of the network device you wish to configure (for example,
eth0 ), and rx-queue with the receive queue you wish to configure (for example, rx-0 ).
Set the value of this file to the value of rps_so ck_fl o w_entri es divided by N, where N is
the number of receive queues on a device. For example, if rps_fl o w_entri es is set to
3276 8 and there are 16 configured receive queues, rps_fl o w_cnt should be set to 20 4 8.
For single-queue devices, the value of rps_fl o w_cnt is the same as the value of
rps_so ck_fl o w_entri es.
59
D ata received from a single sender is not sent to more than one CPU. If the amount of data received
from a single sender is greater than a single CPU can handle, configure a larger frame size to reduce
the number of interrupts and therefore the amount of processing work for the CPU. Alternatively,
consider NIC offload options or faster CPUs.
Consider using numactl or taskset in conjunction with RFS to pin applications to specific cores,
sockets, or NUMA nodes. This can help prevent packets from being processed out of order.
6.3.8. Configuring Accelerat ed RFS

Accelerated RFS boosts the speed of RFS by adding hardware assistance. Like RFS, packets are
forwarded based on the location of the application consuming the packet. Unlike traditional RFS,
however, packets are sent directly to a CPU that is local to the thread consuming the data: either the
CPU that is executing the application, or a CPU local to that CPU in the cache hierarchy.
Accelerated RFS is only available if the following conditions are met:
Accelerated RFS must be supported by the network interface card. Accelerated RFS is supported
by cards that export the nd o _rx_fl o w_steer() netd evi ce function.
ntupl e filtering must be enabled.
Once these conditions are met, CPU to queue mapping is deduced automatically based on
traditional RFS configuration. That is, CPU to queue mapping is deduced based on the IRQ affinities
configured by the driver for each receive queue. Refer to Section 6.3.7, Configuring Receive Flow
Steering (RFS) for details on configuring traditional RFS.
Red Hat recommends using accelerated RFS wherever using RFS is appropriate and the network
interface card supports hardware acceleration.
60
Appendix A. T ool Reference
Appendix A. Tool Reference

This appendix provides a quick reference for the various tools in Red Hat Enterprise Linux 7 that can
be used to tweak performance. See the relevant man page for your tool for complete, up-to-date,
detailed reference material.
A.1. irqbalance
irq b alan ce is a command line tool that distributes hardware interrupts across processors to
improve system performance. It runs as a daemon by default, but can be run once only with the -o nesho t option.
The following parameters are useful for improving performance.
- - p o wert h resh
Sets the number of CPUs that can idle before a CPU is placed into powersave mode. If more
CPUs than the threshold are more than 1 standard deviation below the average softirq
workload and no CPUs are more than one standard deviation above the average, and
have more than one irq assigned to them, a CPU is placed into powersave mode. In
powersave mode, a CPU is not part of irq balancing so that it is not woken unnecessarily.
- - h in t p o licy
D etermines how irq kernel affinity hinting is handled. Valid values are exact (irq affinity
hint is always applied), subset (irq is balanced, but the assigned object is a subset of the
affinity hint), or i g no re (irq affinity hint is ignored completely).
- - p o licyscrip t
D efines the location of a script to execute for each interrupt request, with the device path
and irq number passed as arguments, and a zero exit code expected by irq b alan ce. The
script defined can specify zero or more key value pairs to guide irq b alan ce in managing
the passed irq.
The following are recognized as valid key value pairs.
b an
Valid values are true (exclude the passed irq from balancing) or fal se (perform
balancing on this irq).
b alan ce_level
Allows user override of the balance level of the passed irq. By default the balance
level is based on the PCI device class of the device that owns the irq. Valid values
are no ne, packag e, cache, or co re.
n u ma_n o d e
Allows user override of the NUMA node that is considered local to the passed irq.
If information about the local node is not specified in ACPI, devices are
considered equidistant from all nodes. Valid values are integers (starting from 0)
that identify a specific NUMA node, and -1, which specifies that an irq should be
considered equidistant from all nodes.
- - b an irq
The interrupt with the specified interrupt request number is added to the list of banned
61
The interrupt with the specified interrupt request number is added to the list of banned
interrupts.
You can also use the IRQBALANCE_BANNED_CPUS environment variable to specify a mask of CPUs
that are ignored by irq b alan ce.
For further details, see the man page:
$ man irqbalance
A.2. T una
T u n a allows you to control processor and scheduling affinity. This section covers the command line
interface, but a graphical interface with the same range of functionality is also available. Launch the
graphical utility by running tuna at the command line.
T u n a accepts a number of command line parameters, which are processed in sequence. The
following command distributes load across a four socket system.
tuna --socket 0 --isolate \
--thread my_real_time_app --move \
--irq serial --socket 1 --move \
--irq eth* --socket 2 --spread \
--show_threads --show_irqs
- - g u i
Starts the graphical user interface.
- - cp u s
Takes a comma-delimited list of CPUs to be controlled by T u n a. The list remains in effect
until a new list is specified.
- - co n f ig _f ile_ap p ly
Takes the name of a profile to apply to the system.
- - co n f ig _f ile_list
Lists the pre-loaded profiles.
- - cg ro u p
Used in conjunction with --show_threads. D isplays the type of control group that
processes displayed with --show_threads belong to, if control groups are enabled.
Requires -P
- - af f ect _ch ild ren
When specified, T u n a affects child threads as well as parent threads.
- - f ilt er
D isables the disolay of selected CPUs in --gui. Requires -c
- - iso lat e
Takes a comma-delimited list of CPUs. T u n a migrates all threads away from the CPUs
62
Takes a comma-delimited list of CPUs. T u n a migrates all threads away from the CPUs
specified. Requires -c or -s
- - in clu d e
Takes a comma-delimited list of CPUs. Tuna allows all threads to run on the CPUs
specified. Requires -c or -s
- - n o _kt h read s
When this parameter is specified, Tuna does not affect kernel threads.
- - mo ve
Moves selected entities to CPU-List. Requires -c or -s.
- - p rio rit y
Specifies the scheduler policy and priority for a thread. Valid scheduler policies are O T HER ,
FIFO , R R , BAT C H, or ID LE.
When the policy is FIFO or R R , valid priority values are integers from 1 (lowest) to 99
(highest). The default value is 1. For example, tuna --thread s 786 1 -pri o ri ty= R R : 4 0 sets a policy of R R (round-robin) and a priority of 4 0 for thread 786 1.
When the policy is O T HER , BAT C H, or ID LE, the only valid priority value is 0 , which is also
the default. Requires -t.
- - sh o w_t h read s
Show the thread list.
- - sh o w_irq s
Show the IRQ list.
- - irq s
Takes a comma-delimited list of IRQs that T u n a affects. The list remains in effect until a new
list is specified. IRQs can be added to the list by using + and removed from the list by using
-.
- - save
Saves the kernel threads schedules to the specified file.
- - so cket s
Takes a comma-delimited list of CPU sockets to be controlled by T u n a. This option takes
into account the topology of the system, such as the cores that share a single processor
cache, and that are on the same physical chip.
- - t h read s
Takes a comma-delimited list of threads to be controlled by T u n a. The list remains in effect
until a new list is specified. Threads can be added to the list by using + and removed from
the list by using -.
- - n o _u t h read s
Prevents the operation from affecting user threads.
- - wh at _is
63
To see further help on selected entities. Requires -t

- - sp read
D istribute the threads specified with --thread s evenly between the CPUs specified with -cpus. Requires -t or -q.
- - h elp
Print a list of options. tuna will exit after this action, ignoring the remainder of the commandline.
- - versio n
Shows version.
A.3. et ht ool
The et h t o o l utility allows administrators to view and edit network interface card settings. It is useful
for observing the statistics of certain devices, such as the number of packets dropped by that device.
et h t o o l, its options, and its usage, are comprehensively documented on the man page.
$ man ethtool
A.4 . ss
to assess device performance over time. By default, ss lists open non-listening TCP sockets that
have established connections, but a number of useful options are provided to help administrators
filter out statistics about specific sockets.
One commonly used command is ss -tmpi e, which displays all TCP sockets (t, internal TCP
information (i ), socket memory usage (m), processes using the socket (p), and detailed socket
information (i ).
Red Hat recommends ss over n et st at in Red Hat Enterprise Linux 7.
$ man ss
A.5. t uned
T u n ed is a tuning daemon that can adapt the operating system to perform better under certain
workloads by setting a tuning profile. It can also be configured to react to changes in CPU and
network use and adjusts settings to improve performance in active devices and reduce power
consumption in inactive devices.
To configure dynamic tuning behavior, edit the dynamic_tuning parameter in the
/etc/tuned /tuned -mai n. co nf file. You can also configure the amount of time in seconds
between tuned checking usage and updating tuning details with the update_interval parameter.
For further details about tuned, see the man page:
64
$ man tuned
A.6. t uned-adm
t u n ed - ad m is a command line tool that provides a number of different profiles to improve
performance in a number of specific use cases. It also provides a sub-command (tuned -ad m
reco mmend ) that assesses your system and outputs a recommended tuning profile. This also sets
the default profile for your system at install time, so can be used to return to the default profile.
As of Red Hat Enterprise Linux 7, t u n ed - ad m includes the ability to run any command as part of
enabling or disabling a tuning profile. This allows you to add environment specific checks that are
not available in t u n ed - ad m, such as checking whether the system is the master database node
before selecting which tuning profile to apply.
Red Hat Enterprise Linux 7 also provides the include parameter in profile definition files, allowing
you to base your own t u n ed - ad m profiles on existing profiles.
The following tuning profiles are provided with t u n ed - ad m and are supported in Red Hat
Enterprise Linux 7.
t h ro u g h p u t - p erf o rman ce
A server profile focused on improving throughput. This is the default profile, and is
recommended for most systems.
This profile favors performance over power savings by setting i ntel _pstate and
mi n_perf_pct= 10 0 . It enables transparent huge pages, uses cp u p o wer to set the
perfo rmance cpufreq governor, and sets the input/output scheduler to d ead l i ne. It also
sets kernel.sched_min_granularity_ns to 10 s,
kernel.sched_wakeup_granularity_ns to 15 s, and vm.dirty_ratio to 4 0 % .
l at en cy- p erf o rman ce
A server profile focused on lowering latency. This profile is recommended for latencysensitive workloads that benefit from c-state tuning and the increased TLB efficiency of
transparent huge pages.
max_perf_pct= 10 0 . It enables transparent huge pages, uses cp u p o wer to set the
perfo rmance cpufreq governor, and requests a cpu_dma_latency value of 1.
n et wo rk- lat en cy
A server profile focused on lowering network latency.
mi n_perf_pct= 10 0 . It disables transparent huge pages, and automatic NUMA balancing.
It also uses cp u p o wer to set the perfo rmance cpufreq governor, and requests a
cpu_dma_latency value of 1. It also sets busy_read and busy_poll times to 50 s,
and tcp_fastopen to 3.
n et wo rk- t h ro u g h p u t
A server profile focused on improving network throughput.
max_perf_pct= 10 0 and increasing kernel network buffer sizes. It enables transparent
huge pages, and uses cp u p o wer to set the perfo rmance cpufreq governor. It also sets
65
kernel.sched_min_granularity_ns to 10 s,
virt u al- g u est
A profile focused on optimizing performance in Red Hat Enterprise Linux 7 virtual machines.
max_perf_pct= 10 0 . It also decreases the swappiness of virtual memory. It enables
transparent huge pages, and uses cp u p o wer to set the perfo rmance cpufreq governor. It
also sets kernel.sched_min_granularity_ns to 10 s,
virt u al- h o st
A profile focused on optimizing performance in Red Hat Enterprise Linux 7 virtualization
hosts.
max_perf_pct= 10 0 . It also decreases the swappiness of virtual memory. This profile
enables transparent huge pages and writes dirty pages back to disk more frequently. It
uses cp u p o wer to set the perfo rmance cpufreq governor. It also sets
kernel.sched_min_granularity_ns to 10 s,
kernel.sched_wakeup_granularity_ns to 15 s, kernel.sched_migration_cost
to 5 s, and vm.dirty_ratio to 4 0 % .
For detailed information about the power saving profiles provided with tuned-adm, see the Red Hat
Enterprise Linux 7 Power Management Guide, available from
For detailed information about using t u n ed - ad m, see the man page:
$ man tuned-adm
A.7. perf
The p erf tool provides a number of useful commands, some of which are listed in this section. For
detailed information about p erf , see the Red Hat Enterprise Linux 7 Developer Guide, available from
http://access.redhat.com/site/documentation/Red_Hat_Enterprise_Linux/, or refer to the man pages.
p erf st at
This command provides overall statistics for common performance events, including
instructions executed and clock cycles consumed. You can use the option flags to gather
statistics on events other than the default measurement events. As of Red Hat
Enterprise Linux 6.4, it is possible to use perf stat to filter monitoring based on one or
more specified control groups (cgroups).
For further information, read the man page:
$ man perf-stat
p erf reco rd
This command records performance data into a file which can be later analyzed using
perf repo rt. For further details, read the man page:
66
$ man perf-record
p erf rep o rt
This command reads the performance data from a file and analyzes the recorded data. For
further details, read the man page:
$ man perf-report
p erf list
This command lists the events available on a particular machine. These events vary based
on the performance monitoring hardware and the software configuration of the system. For
further information, read the man page:
$ man perf-list
p erf t o p
This command performs a similar function to the t o p tool. It generates and displays a
performance counter profile in realtime. For further information, read the man page:
$ man perf-top
p erf t race
This command performs a similar function to the st race tool. It monitors the system calls
used by a specified thread or process and all signals received by that application.
Additional trace targets are available; refer to the man page for a full list:
$ man perf-trace
A.8. Performance Co-Pilot (PCP)

Performance Co-Pilot (PCP) provides a large number of command line tools, graphical tools, and
libraries. For more information about any of these tools, see the man page: type man toolname at
the command line, replacing toolname with the name of the tool.
The pcp-doc package installs detailed documentation to the /usr/share/d o c/pcp-d o c directory
by default.
The Red Hat Customer Portal has a number of useful articles to help you get started with PCP. An
index of these articles is available here: https://access.redhat.com/articles/1145953.
A.9. vmst at
Vmst at outputs reports on your system's processes, memory, paging, block input/output, interrupts,
and CPU activity. It provides an instantaneous report of the average of these events since the
machine was last booted, or since the previous report.
- a
D isplays active and inactive memory.
67
- f
D isplays the number of forks since boot. This includes the fo rk, vfo rk, and cl o ne
system calls, and is equivalent to the total number of tasks created. Each process is
represented by one or more tasks, depending on thread usage. This display does not
repeat.
- m
D isplays slab information.
- n
Specifies that the header will appear once, not periodically.
- s
D isplays a table of various event counters and memory statistics. This display does not
repeat.
d elay
The delay between reports in seconds. If no delay is specified, only one report is printed,
with the average values since the machine was last booted.
c o u n t
The number of times to report on the system. If no count is specified and delay is defined,
vmst at reports indefinitely.
- d
D isplays disk statistics.
- p
Takes a partition name as a value, and reports detailed statistics for that partition.
- S
D efines the units output by the report. Valid values are k (1000 bytes), K (1024 bytes), m
(1000000 bytes), or M (1048576 bytes).
- D
Report summary statistics about disk activity.
For detailed information about the output provided by each output mode, see the man page:
$ man vmstat
A.10. x86_energy_perf_policy
The x86 _en erg y_p erf _p o licy tool allows administrators to define the relative importance of
performance and energy efficiency. It is provided by the kernel-tools package.
To view the current policy, run the following command:
# x86_energy_perf_policy -r
68
To set a new policy, run the following command:

# x86_energy_perf_policy profile_name
Replace profile_name with one of the following profiles.
p erf o rman ce
The processor does not sacrifice performance for the sake of saving energy. This is the
default value.
n o rmal
The processor tolerates minor performance compromises for potentially significant energy
savings. This is a reasonable saving for most servers and desktops.
p o wersave
The processor accepts potentially significant performance decreases in order to maximize
energy efficiency.
For further details of how to use x86 _en erg y_p erf _p o licy, see the man page:
$ man x86_energy_perf_policy
A.11. t urbost at
The t u rb o st at tool provides detailed information about the amount of time that the system spends in
different states. T u rb o st at is provided by the kernel-tools package.
By default, t u rb o st at prints a summary of counter results for the entire system, followed by counter
results every 5 seconds, under the following headings:
p kg
The processor package number.
c o re
The processor core number.
C PU
The Linux CPU (logical processor) number.
% c0
The percentage of the interval for which the CPU retired instructions.
G H z
When this number is higher than the value in TSC, the CPU is in turbo mode
T SC
The average clock speed over the course of the entire interval.
% c1, % c3, an d % c6
69
The percentage of the interval for which the processor was in the c1, c3, or c6 state,
respectively.
% p c3 o r % p c6
The percentage of the interval for which the processor was in the pc3 or pc6 state,
respectively.
Specify a different period between counter results with the -i option, for example, run turbo stat i 10 to print results every 10 seconds instead.
Note
Upcoming Intel processors may add additional c-states. As of Red Hat Enterprise Linux 7.0,
t u rb o st at provides support for the c7, c8, c9, and c10 states.
A.12. numast at
The n u mast at tool is provided by the numactl package, and displays memory statistics (such as
allocation hits and misses) for processes and the operating system on a per-NUMA-node basis. The
default tracking categories for the numastat command are outlined as follows:
n u ma_h it
The number of pages that were successfully allocated to this node.
n u ma_miss
The number of pages that were allocated on this node because of low memory on the
intended node. Each numa_mi ss event has a corresponding numa_fo rei g n event on
another node.
n u ma_f o reig n
The number of pages initially intended for this node that were allocated to another node
instead. Each numa_fo rei g n event has a corresponding numa_mi ss event on another
node.
i n t erleave_h it
The number of interleave policy pages successfully allocated to this node.
l o cal_n o d e
The number of pages successfully allocated on this node, by a process on this node.
o t h er_n o d e
The number of pages allocated on this node, by a process on another node.
Supplying any of the following options changes the displayed units to megabytes of memory
(rounded to two decimal places), and changes other specific n u mast at behaviors as described
below.
- c
70
Horizontally condenses the displayed table of information. This is useful on systems with a
large number of NUMA nodes, but column width and inter-column spacing are somewhat
unpredictable. When this option is used, the amount of memory is rounded to the nearest
megabyte.
- m
D isplays system-wide memory usage information on a per-node basis, similar to the
information found in /pro c/memi nfo .
- n
D isplays the same information as the original numastat command (numa_hi t,
numa_mi ss, numa_fo rei g n, i nterl eave_hi t, l o cal _no d e, and o ther_no d e), with
an updated format, using megabytes as the unit of measurement.
- p p at t ern
D isplays per-node memory information for the specified pattern. If the value for pattern is
comprised of digits, n u mast at assumes that it is a numerical process identifier. Otherwise,
n u mast at searches process command lines for the specified pattern.
Command line arguments entered after the value of the -p option are assumed to be
additional patterns for which to filter. Additional patterns expand, rather than narrow, the
filter.
- s
Sorts the displayed data in descending order so that the biggest memory consumers
(according to the total column) are listed first.
Optionally, you can specify a node, and the table will be sorted according to the node
column. When using this option, the node value must follow the -s option immediately, as
shown here:
numastat -s2
D o not include white space between the option and its value.
- v
D isplays more verbose information. Namely, process information for multiple processes will
display detailed information for each process.
- V
D isplays numastat version information.
- z
Omits table rows and columns with only zero values from the displayed information. Note
that some near-zero values that are rounded to zero for display purposes will not be
omitted from the displayed output.
A.13. numact l
71
N u mact l lets administrators run a process with a specified scheduling or memory placement policy.
N u mact l can also set a persistent policy for shared memory segments or files, and set the processor
affinity and memory affinity of a process.
N u mact l provides a number of useful options. This appendix outlines some of these options and
gives suggestions for their use, but is not exhaustive.
- - h ard ware
D isplays an inventory of available nodes on the system, including relative distances
between nodes.
- - memb in d
Ensures that memory is allocated only from specific nodes. If there is insufficient memory
available in the specified location, allocation fails.
- - cp u n o d eb in d
Ensures that a specified command and its child processes execute only on the specified
node.
- - p h ycp u b in d
Ensures that a specified command and its child processes execute only on the specified
processor.
- - lo calallo c
Specifies that memory should always be allocated from the local node.
- - p ref erred
Specifies a preferred node from which to allocate memory. If memory cannot be allocated
from this specified node, another node will be used as a fallback.
For further details about these and other parameters, see the man page:
$ man numactl
A.14 . numad
n u mad is an automatic NUMA affinity management daemon. It monitors NUMA topology and
resource usage within a system in order to dynamically improve NUMA resource allocation and
management.
Note that when n u mad is enabled, its behavior overrides the default behavior of automatic NUMA
balancing.
A.14 .1. Using numad from t he command line

To use n u mad as an executable, just run:
# numad
While n u mad runs, its activities are logged in /var/l o g /numad . l o g . It will run until stopped with
the following command:
72
# numad -i 0
Stopping n u mad does not remove the changes it has made to improve NUMA affinity. If system use
changes significantly, running n u mad again will adjust affinity to improve performance under the
new conditions.
To restrict n u mad management to a specific process, start it with the following options.
# numad -S 0 -p pid
- p pid
This option adds the specified pid to an explicit inclusion list. The process specified will not
be managed until it meets the n u mad process significance threshold.
- S 0
This sets the type of process scanning to 0 , which limits n u mad management to explicitly
included processes.
For further information about available n u mad options, refer to the n u mad man page:
$ man numad
A.14 .2. Using numad as a service

While n u mad runs as a service, it attempts to tune the system dynamically based on the current
system workload. Its activities are logged in /var/l o g /numad . l o g .
To start the service, run:
# systemctl start numad.service
To make the service persist across reboots, run:
# chkconfig numad on
For further information about available n u mad options, refer to the n u mad man page:
$ man numad
A.14 .3. Pre-placement advice

n u mad provides a pre-placement advice service that can be queried by various job management
systems to provide assistance with the initial binding of CPU and memory resources for their
processes. This pre-placement advice is available regardless of whether n u mad is running as an
executable or a service.
A.14 .4 . Using numad wit h KSM

If KSM is in use on a NUMA system, change the value of the /sys/kernel /mm/ksm/merg e_no d es
parameter to 0 to avoid merging pages across NUMA nodes. Otherwise, KSM increases remote
memory accesses as it merges pages across nodes. Furthermore, kernel memory accounting
73
statistics can eventually contradict each other after large amounts of cross-node merging. As such,
numad can become confused about the correct amounts and locations of available memory, after the
KSM daemon merges many memory pages. KSM is beneficial only if you are overcommitting the
memory on your system. If your system has sufficient free memory, you may achieve higher
performance by turning off and disabling the KSM daemon.
A.15. OProfile
OProfile is a low overhead, system-wide performance monitoring tool provided by the oprofile
package. It uses the performance monitoring hardware on the processor to retrieve information about
the kernel and executables on the system, such as when memory is referenced, the number of
second-level cache requests, and the number of hardware interrupts received. OProfile is also able to
profile applications that run in a Java Virtual Machine (JVM).
OProfile provides the following tools. Note that the legacy o pco ntro l tool and the new o perf tool
are mutually exclusive.
o p h elp
D isplays available events for the systems processor along with a brief description of each.
o p imp o rt
Converts sample database files from a foreign binary format to the native format for the
system. Only use this option when analyzing a sample database from a different
architecture.
o p an n o t at e
Creates annotated source for an executable if the application was compiled with debugging
symbols.
o p co n t ro l
Configures which data is collected in a profiling run.
o p erf
Intended to replace o pco ntro l . The o perf tool uses the Linux Performance Events
subsystem, allowing you to target your profiling more precisely, as a single process or
system-wide, and allowing OProfile to co-exist better with other tools using the performance
monitoring hardware on your system. Unlike o pco ntro l , no initial setup is required, and it
can be used without the root privileges unless the --system-wi d e option is in use.
o p rep o rt
Retrieves profile data.
o p ro f iled
Runs as a daemon to periodically write sample data to disk.
Legacy mode (o pco ntro l , o pro fi l ed , and post-processing tools) remains available, but is no
longer the recommended profiling method.
For further information about any of these commands, see the OProfile man page:
$ man oprofile
74
A.16. t askset
The t askset tool is provided by the util-linux package. It allows administrators to retrieve and set the
processor affinity of a running process, or launch a process with a specified processor affinity.
Important
t askset does not guarantee local memory allocation. If you require the additional
performance benefits of local memory allocation, Red Hat recommends using n u mact l
instead of taskset.
To set the CPU affinity of a running process, run the following command:
# taskset -c processors pid
Replace processors with a comma delimited list of processors or ranges of processors (for example,
1,3,5-7. Replace pid with the process identifier of the process that you want to reconfigure.
To launch a process with a specified affinity, run the following command:
# taskset -c processors -- application
Replace processors with a comma delimited list of processors or ranges of processors. Replace
application with the command, options and arguments of the application you want to run.
For more information about t askset , see the man page:
$ man taskset
A.17. Syst emT ap

Syst emT ap is extensively documented in its own guides. The Red Hat Enterprise Linux 7 versions of
the SystemTap Beginner's Guide and the SystemTap TapSet Reference are both available from
75
Appendix B. Revision History

R evisio n 10.08- 4 0
Wed Au g 21 2015
Rebuilt guide for RHEL 7.2 Beta
C h arlie B o yle
R evisio n 10.08- 35
Wed Au g 6 2015
Added BTRFS information to 5.3.7.3
C h arlie B o yle
R evisio n 04 .7- 33
Wed Au g 5 2015
Added BTRFS compression information to 5.1.3
C h arlie B o yle
R evisio n 03.6 - 32
Mo n Au g 3 2015
Changed heading level for file systems in 5.1.3 File Systems
C h arlie B o yle
Wed Ju n 5 2015
Updated SSD information 5.5.1 part of 1189350.
C h arlie B o yle
Wed Ap r 8 2015
C h arlie B o yle
Updated VMSTAT parameters in 5.2.1 to reflect man page, BZ 1131829.
Wed Ap r 8 2015
Updated VMSTAT parameters to reflect man page, BZ 1131829.
C h arlie B o yle
T u e Ap r 7 2015
updated Tuna help, BZ 1131829.
C h arlie B o yle
T u e Ap r 7 2015
Changed Turbostat Ghz description, BZ 992461.
C h arlie B o yle
Fri Mar 27 2015
Corrected parameter setting for, BZ 1145906.
C h arlie B o yle
T h u Mar 26 2015
C h arlie B o yle
Corrected details of Section 4.3 Configuration tools, BZ 1122400.
R evisio n 0.3- 25
T h u Mar 26 2015
Corrected details of throughput performance, BZ 1122132.
C h arlie B o yle
R evisio n 0.3- 24
Mo n Feb 23 2015
Corrected details of confirming busy poll support, BZ 1080703.
Lau ra B ailey
R evisio n 0.3- 23
Building for RHEL 7.1 GA.
T u e Feb 17 2015
Lau ra B ailey
R evisio n 0.3- 22
T u e Feb 17 2015
Added busy poll support check, BZ 1080703.
Lau ra B ailey
R evisio n 0.3- 21
T h u Feb 05 2015
Lau ra B ailey
Noted new tuned profile parameter, cmdline, BZ 1130818.
Added note to the 'New in 7.1' section re. new discard parameter in swapon command, BZ 1130826.
76
Appendix B. Revision Hist ory
R evisio n 0.3- 20
Fri Jan 09 2015
Fix error in pgrep command. BZ 1155253.
Improve the description of vm.swappiness. BZ 1148419.
C h arles B o yle
R evisio n 0.3- 19
Fri D ec 05 2014
Updating sort_order for splash page presentation.
Lau ra B ailey
R evisio n 0.3- 18
Wed N o v 26 2014
Added link to PCP article index for 7.1 Beta, BZ 1083387.
Lau ra B ailey
R evisio n 0.3- 15
Mo n N o v 24 2014
Added note on idle balancing changes, BZ 1131851.
Lau ra B ailey
R evisio n 0.3- 14
Wed N o v 19 2014
Lau ra B ailey
Updated huge page allocation details, BZ 1131367.
Added example of how to allocate huge pages during runtime, BZ 1131367.
R evisio n 0.3- 12
T h u N o v 13 2014
Lau ra B ailey
Added new default values for SHMALL and SHMMAX in RHEL 7.1, BZ 1131848.
R evisio n 0.3- 11
Mo n N o v 10 2014
Lau ra B ailey
Added note about support status of clustered allocation / bigalloc for ext4, BZ 794607.
R evisio n 0.3- 10
Fri O ct 31 2014
D ocumented per-node static huge pages, BZ 1131832.
Lau ra B ailey
R evisio n 0.3- 9
T u e Ju l 22 2014
Added description of latencytap.stp script, BZ 988155.
Lau ra B ailey
R evisio n 0.3- 7
T h u Ju n 26 2014
Lau ra B ailey
Corrected typographical error in the CPU chapter; thanks Jiri Hladky.
Removed references to tuned altering the I/O scheduler; thanks Jiri Hladky.
R evisio n 0.3- 5
Wed Ju n 11 2014
Lau ra B ailey
Added trailing slash to access.redhat.com links that wouldn't redirect.
R evisio n 0.3- 4
T u e Ju n 10 2014
Lau ra B ailey
Added interrupt and CPU banning details to irqbalance appendix BZ 852981.
R evisio n 0.3- 3
Rebuilding for RHEL 7.0 GA.
Mo n Ap r 07 2014
Lau ra B ailey
R evisio n 0.3- 2
Mo n Ap r 07 2014
Updated book structure for RT#294949.
Lau ra B ailey
R evisio n 0.2- 38
Mo n Ap r 07 2014
Added updated OProfile data, BZ 955882.
Removing outdated comments.
Lau ra B ailey
R evisio n 0.2- 34
Lau ra B ailey
Fri Ap r 04 2014
77
Corrected lstopo output image styling, BZ 1042800.

Added details about irqbalance daemon, BZ 955890.
Added and corrected details about control groups, BZ 794624.
Added details about PCP, BZ 955883.
Updated XFS tuning details, BZ 794616.
Added updated OProfile data, BZ 955882.
R evisio n 0.2- 27
Fri Mar 28 2014
Lau ra B ailey
Corrected busy_poll section based on feedback from Jeremy Eder, RT276607.
Corrected nohz_full section and added details based on feedback from Jeremy Eder, RT284423.
Added further detail to SystemTap sections, BZ 955884.
Added further detail to the SSD section, BZ 955900.
Added further detail on the tuned-adm recommend command, BZ 794623.
Corrected note about automatic NUMA balancing in features section, BZ 794612.
Corrected a number of terminology issues and example output issues regarding NUMA, including a
new image, BZ 1042800.
Corrected details about irqbalance in conjunction with RSS based on feedback from Jeremy Eder.
R evisio n 0.2- 19
Fri Mar 21 2014
Lau ra B ailey
Added details about transparent huge pages to the Memory chapter, BZ 794621.
Corrected use of terms related to NUMA nodes, BZ 1042800.
Updated kernel limits, BZ 955894.
D rafted tickless kernel section, RT284423.
D rafted busy polling section, RT276607.
Updated information about file system barriers.
Removed unclear information about per-node huge page assignment. BZ 1079079 created to add
more useful information in future.
Added details about solid state disks, BZ 955900.
Removed review markers.
R evisio n 0.2- 14
T h u Mar 13 2014
Lau ra B ailey
Applied feedback from Jeremy Eder and Joe Mario.
Noted updates to Tuna GUI from BZ 955872.
Added details about SystemTap to the Networking chapter and Tools Reference appendix,
BZ 955884.
R evisio n 0.2- 12
Fri Mar 07 2014
Lau ra B ailey
Noted support for automatic NUMA migration, as per BZ 794612.
Applied additional feedback from Jeremy Eder.
R evisio n 0.2- 11
Fri Mar 07 2014
Applied feedback from Jeremy Eder.
Lau ra B ailey
R evisio n 0.2- 10
Mo n Feb 24 2014
Lau ra B ailey
Corrected Ext4 information based on feedback from Luk Czerner (BZ #794607).
R evisio n 0.2- 9
Mo n Feb 17 2014
Lau ra B ailey
Corrected the CPU chapter based on feedback from Bill Gray.
Corrected and added to the Memory chapter and Tools Reference based on feedback from Bill Gray.
R evisio n 0.2- 8
78
Mo n Feb 10 2014
Lau ra B ailey
Appendix B. Revision Hist ory
Added isolcpus boot parameter to CPU chapter (RT276607).

SME feedback: corrected parameter descriptions and added new parameters (BZ #970844).
Added recommended tuned-adm profiles to Network chapter.
Added remarks to flag sections for review.
R evisio n 0.2- 4
Mo n Feb 03 2014
Lau ra B ailey
Confirmed that numactl --membi nd parameter is documented (BZ #922070).
Added details about Tuna to the Tools introduction, the CPU chapter, and the Tools Reference
appendix (BZ #970844).
Corrected structural error in the Storage and File Systems chapter.
Added missing cross references.
R evisio n 0.2- 2
Fri Jan 31 2014
Lau ra B ailey
Rewrite and restructure complete.
Ensured that all tools mentioned in the guide were listed alongside the package that provides them.
R evisio n 0.1- 11
T h u D ec 05 2013
Building restructured guide for RHEL 7.0 Beta.
Lau ra B ailey
R evisio n 0.1- 10
Pre-Beta customer build.
Lau ra B ailey
Wed N o v 27 2013
R evisio n 0.1- 9
T u e O ct 15 2013
Minor corrections based on customer feedback (BZ #1011676).
Lau ra B ailey
R evisio n 0.1- 7
Mo n Sep 09 2013
Merged new content from RHEL 6.5.
Applied editor feedback.
Lau ra B ailey
R evisio n 0.1- 6
Wed May 29 2013
Lau ra B ailey
Updated ext4 file system limits (BZ #794607).
Corrected theoretical maximum of a 64-bit file system.
Added the New Features section to track performance-related changes.
Changed default I/O scheduler from cfq to d ead l i ne (BZ #794602).
Added draft content for BTRFS tuning (BZ #794604).
Updated XFS section to provide clearer recommendations about directory block sizes, and updated
XFS supported limits (BZ #794616.
R evisio n 0.1- 2
T h u rs Jan 31 2013
Updated and published as RHEL 7 draft.
T ah lia R ich ard so n
R evisio n 0.1- 1
Wed Jan 16 2013
Branched from the RHEL 6.4 version of this document.
Lau ra B ailey
79

Red Hat Enterprise Linux 7 Beta Performance Tuning Guide en US

Uploaded by

Copyright:

Available Formats

Red Hat Enterprise Linux 7 Beta Performance Tuning Guide en US

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Red Hat Enterprise Linux 7 Beta Performance Tuning Guide en US

Uploaded by

Copyright:

Available Formats

Red Hat Enterprise Linux 7.

Optimizing subsystem throughput in Red Hat Enterprise Linux 7

Red Hat Subject Matter Experts

Red Hat Enterprise Linux 7.2 Beta Performance Tuning Guide

Optimizing subsystem throughput in Red Hat Enterprise Linux 7

T able of Cont ent s

2 .15. O Pro file

3 .3. Co nfig uratio n s ug g es tio ns

5 .3. Co nfig uratio n to o ls

Performance T uning G uide

Chapt er 1 . Performance Feat ures in Red Hat Ent erprise Linux 7

Chapter 1. Performance Features in Red Hat Enterprise Linux 7

1.1. New in 7.1

1.2. New in 7.0

Performance T uning G uide

Chapt er 2 . Performance Monit oring T ools

Chapter 2. Performance Monitoring Tools

2.2. GNOME Syst em Monit or

Performance T uning G uide

2.3. Performance Co-Pilot (PCP)

2.5. Built in command line t ools

Chapt er 2 . Performance Monit oring T ools

2.5.3. Virt ual Memory St at ist ics (vmst at )

2.5.4 . Syst em Act ivit y Report er (sar)

2.6. t uned and t uned-adm

Performance T uning G uide

low latency for storage and network

Chapt er 2 . Performance Monit oring T ools

Performance T uning G uide

2.14 . Syst emT ap

Chapt er 2 . Performance Monit oring T ools

Performance T uning G uide

3.1. Considerat ions

3.1.1. Syst em T opology

processors, memory, and peripheral devices are connected. The

3.1 .1 .1 . De t e rm ining syst e m t o po lo gy

Performance T uning G uide

Output of lstopo command

Performance T uning G uide

3.1 .2 .1 . Ke rne l T icks

3.1.3. Int errupt Request (IRQ) Handling

3.2. Monit oring and diagnosing performance problems

3.2.3. /proc/int errupt s

Performance T uning G uide

If a particular application or device is generating a large number of interrupt requests to be handled

3.3. Configurat ion suggest ions

3.3.1. Configuring kernel t ick t ime

# perf stat -C 1 -e irq_vectors:local_timer_entry taskset -c 1 stress -t

3.3.2. Set t ing hardware performance policy (x86_energy_perf_policy)

3.3.3. Set t ing process affinit y wit h t askset

3.3.4 . Managing NUMA affinit y wit h numact l

Performance T uning G uide

3.3.5. Aut omat ic NUMA affinit y management wit h numad

3.3.6. T uning scheduling policy

3.3.6 .1 . Sche duling po licie s

Performance T uning G uide

3.3.6 .2 . Iso lat ing CPUs

3.3.7. Set t ing int errupt affinit y

Performance T uning G uide

4 .1. Considerat ions