feat(bigquery/storage/managedwriter): support append retries #6695
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
yirutang
reviewed
Sep 22, 2022
yirutang
approved these changes
Sep 27, 2022
yirutang
reviewed
Sep 27, 2022
| // Any other non-status based errors treated as retryable. | ||
| return r.bo.Pause(), true | ||
| // All other non-status errors are treated as non-retryable (including context errors). | ||
| return | ||
| } | ||
| switch s.Code() { | ||
| case codes.Aborted, | ||
| codes.Canceled, | ||
| codes.DeadlineExceeded, | ||
| codes.Internal, | ||
| codes.Unavailable: |
There was a problem hiding this comment.
Sorry, missed this, please retry on failed_precondition.
kimihrr
pushed a commit
to kimihrr/google-cloud-go-kimia
that referenced
this pull request
Sep 28, 2022
…pis#6695) This PR adds support for automatic retry of failed appends. Failures are evaluated both from the perspective of any receive errors getting the response, as well as any response that may be embedded in response from the service. Previous behavior was to simply re-attempt failures when issuing the append. This PR also adds a new WriterOption (`DisableWriteRetries(disable bool)`) to control this behavior (default is to have retries enabled). For users sensitive to write duplication, this PR also exposes a new TotalAttempts method on the AppendResult, which will indicate the total number of times this write was attempted. This also tries to clean up retries in general a bit more. The generated client will already retry unary RPCs, subject to the [service config](https://github.com/googleapis/googleapis/blob/master/google/cloud/bigquery/storage/v1/bigquerystorage_grpc_service_config.json) present when generating the storage API. We specifically clarify/introduce two additional kinds of retries above that: a unary retry and a stateless retry. The unary retry is used to (re)open the underlying bidi network connection which appends are sent upon, as we want to be resilient to reconnection. The unary retry is effectively stateful for the operation of reopening the connection, and thus uses a gax-based backoff that backs off with increasing intervals. The stateless retry is used when processing the responses returning from the backend on the bidi stream, where backing off incrementally doesn't make sense. Instead, we use a base backoff and jitter, and for cases where a more severe backoff is warranted (throughput exhaustion) we use a multiplication factor. The intent here is to provide backpressure which will eventually saturate the append queue and cause blocking/rejection of writes until the backlog recovers.
gcf-merge-on-green bot
pushed a commit
that referenced
this pull request
Oct 24, 2022
🤖 I have created a release *beep* *boop* --- ## [1.43.0](https://togithub.com/googleapis/google-cloud-go/compare/bigquery/v1.42.0...bigquery/v1.43.0) (2022-10-24) ### Features * **bigquery/analyticshub:** rename nodejs analyticshub library package name ([52dddd1](https://togithub.com/googleapis/google-cloud-go/commit/52dddd1ed89fbe77e1859311c3b993a77a82bfc7)) * **bigquery/storage/managedwriter:** Enable field name indirection ([#6247](https://togithub.com/googleapis/google-cloud-go/issues/6247)) ([1969273](https://togithub.com/googleapis/google-cloud-go/commit/19692735b0fbafa176d0315bda923528e1eedf6d)) * **bigquery/storage/managedwriter:** Retry on FailedPrecondition ([#6761](https://togithub.com/googleapis/google-cloud-go/issues/6761)) ([d1a444d](https://togithub.com/googleapis/google-cloud-go/commit/d1a444d769c9578b586bef608d343b4b0abd3658)) * **bigquery/storage/managedwriter:** Support append retries ([#6695](https://togithub.com/googleapis/google-cloud-go/issues/6695)) ([6ae9c67](https://togithub.com/googleapis/google-cloud-go/commit/6ae9c670a11d80b34872cb05fda933303b81851d)) * **bigquery/storage/managedwriter:** Switch to opt-in retry ([#6765](https://togithub.com/googleapis/google-cloud-go/issues/6765)) ([a3e97a6](https://togithub.com/googleapis/google-cloud-go/commit/a3e97a6f15ad1989ef815b9bd5838192f9f226f1)) * **bigquery:** Add remote function options to routine metadata ([#6702](https://togithub.com/googleapis/google-cloud-go/issues/6702)) ([d9a437d](https://togithub.com/googleapis/google-cloud-go/commit/d9a437de75a5f5151cd000d8f9a6b7fc567d8551)) ### Bug Fixes * **bigquery/storage/managedwriter:** Address possible resource leak ([#6775](https://togithub.com/googleapis/google-cloud-go/issues/6775)) ([979440b](https://togithub.com/googleapis/google-cloud-go/commit/979440b43573c1cfd744d3f63d0c633c7319ad46)) * **bigquery:** Avoid stack overflow on query param with recursive types ([#6890](https://togithub.com/googleapis/google-cloud-go/issues/6890)) ([854ccfc](https://togithub.com/googleapis/google-cloud-go/commit/854ccfca259d747759d283fc0d0053893f3c8f8d)), refs [#6884](https://togithub.com/googleapis/google-cloud-go/issues/6884) * **bigquery:** Bq connection auth scopes ([#6752](https://togithub.com/googleapis/google-cloud-go/issues/6752)) ([8e09288](https://togithub.com/googleapis/google-cloud-go/commit/8e09288185f721d90288c3aa873980fc44b98613)), refs [#6744](https://togithub.com/googleapis/google-cloud-go/issues/6744) ### Documentation * **bigquery/storage/managedwriter/adapt:** Typo in error string ([#6729](https://togithub.com/googleapis/google-cloud-go/issues/6729)) ([bb26153](https://togithub.com/googleapis/google-cloud-go/commit/bb26153d38475cd9784edbf241df84c368f5a166)) * **bigquery/storage/managedwriter:** Add retry info to package docs ([#6803](https://togithub.com/googleapis/google-cloud-go/issues/6803)) ([81e52e5](https://togithub.com/googleapis/google-cloud-go/commit/81e52e59dcf3c4a44108e039fb0e3a0e2ce8284f)) --- This PR was generated with [Release Please](https://togithub.com/googleapis/release-please). See [documentation](https://togithub.com/googleapis/release-please#release-please).
kimihrr
pushed a commit
to kimihrr/google-cloud-go-kimia
that referenced
this pull request
Oct 25, 2022
🤖 I have created a release *beep* *boop* --- ## [1.43.0](https://togithub.com/googleapis/google-cloud-go/compare/bigquery/v1.42.0...bigquery/v1.43.0) (2022-10-24) ### Features * **bigquery/analyticshub:** rename nodejs analyticshub library package name ([52dddd1](https://togithub.com/googleapis/google-cloud-go/commit/52dddd1ed89fbe77e1859311c3b993a77a82bfc7)) * **bigquery/storage/managedwriter:** Enable field name indirection ([googleapis#6247](https://togithub.com/googleapis/google-cloud-go/issues/6247)) ([1969273](https://togithub.com/googleapis/google-cloud-go/commit/19692735b0fbafa176d0315bda923528e1eedf6d)) * **bigquery/storage/managedwriter:** Retry on FailedPrecondition ([googleapis#6761](https://togithub.com/googleapis/google-cloud-go/issues/6761)) ([d1a444d](https://togithub.com/googleapis/google-cloud-go/commit/d1a444d769c9578b586bef608d343b4b0abd3658)) * **bigquery/storage/managedwriter:** Support append retries ([googleapis#6695](https://togithub.com/googleapis/google-cloud-go/issues/6695)) ([6ae9c67](https://togithub.com/googleapis/google-cloud-go/commit/6ae9c670a11d80b34872cb05fda933303b81851d)) * **bigquery/storage/managedwriter:** Switch to opt-in retry ([googleapis#6765](https://togithub.com/googleapis/google-cloud-go/issues/6765)) ([a3e97a6](https://togithub.com/googleapis/google-cloud-go/commit/a3e97a6f15ad1989ef815b9bd5838192f9f226f1)) * **bigquery:** Add remote function options to routine metadata ([googleapis#6702](https://togithub.com/googleapis/google-cloud-go/issues/6702)) ([d9a437d](https://togithub.com/googleapis/google-cloud-go/commit/d9a437de75a5f5151cd000d8f9a6b7fc567d8551)) ### Bug Fixes * **bigquery/storage/managedwriter:** Address possible resource leak ([googleapis#6775](https://togithub.com/googleapis/google-cloud-go/issues/6775)) ([979440b](https://togithub.com/googleapis/google-cloud-go/commit/979440b43573c1cfd744d3f63d0c633c7319ad46)) * **bigquery:** Avoid stack overflow on query param with recursive types ([googleapis#6890](https://togithub.com/googleapis/google-cloud-go/issues/6890)) ([854ccfc](https://togithub.com/googleapis/google-cloud-go/commit/854ccfca259d747759d283fc0d0053893f3c8f8d)), refs [googleapis#6884](https://togithub.com/googleapis/google-cloud-go/issues/6884) * **bigquery:** Bq connection auth scopes ([googleapis#6752](https://togithub.com/googleapis/google-cloud-go/issues/6752)) ([8e09288](https://togithub.com/googleapis/google-cloud-go/commit/8e09288185f721d90288c3aa873980fc44b98613)), refs [googleapis#6744](https://togithub.com/googleapis/google-cloud-go/issues/6744) ### Documentation * **bigquery/storage/managedwriter/adapt:** Typo in error string ([googleapis#6729](https://togithub.com/googleapis/google-cloud-go/issues/6729)) ([bb26153](https://togithub.com/googleapis/google-cloud-go/commit/bb26153d38475cd9784edbf241df84c368f5a166)) * **bigquery/storage/managedwriter:** Add retry info to package docs ([googleapis#6803](https://togithub.com/googleapis/google-cloud-go/issues/6803)) ([81e52e5](https://togithub.com/googleapis/google-cloud-go/commit/81e52e59dcf3c4a44108e039fb0e3a0e2ce8284f)) --- This PR was generated with [Release Please](https://togithub.com/googleapis/release-please). See [documentation](https://togithub.com/googleapis/release-please#release-please).
This was referenced Nov 1, 2022
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds support for automatic retry of failed appends. Failures are evaluated both from the perspective of any receive errors getting the response, as well as any response that may be embedded in response from the service. Previous behavior was to simply re-attempt failures when issuing the append.
This PR also adds a new WriterOption (
DisableWriteRetries(disable bool)) to control this behavior (default is to have retries enabled).For users sensitive to write duplication, this PR also exposes a new TotalAttempts method on the AppendResult, which will indicate the total number of times this write was attempted.
This also tries to clean up retries in general a bit more. The generated client will already retry unary RPCs, subject to the service config present when generating the storage API.
We specifically clarify/introduce two additional kinds of retries above that: a unary retry and a stateless retry.
The unary retry is used to (re)open the underlying bidi network connection which appends are sent upon, as we want to be resilient to reconnection. The unary retry is effectively stateful for the operation of reopening the connection, and thus uses a gax-based backoff that backs off with increasing intervals.
The stateless retry is used when processing the responses returning from the backend on the bidi stream, where backing off incrementally doesn't make sense. Instead, we use a base backoff and jitter, and for cases where a more severe backoff is warranted (throughput exhaustion) we use a multiplication factor. The intent here is to provide backpressure which will eventually saturate the append queue and cause blocking/rejection of writes until the backlog recovers.