S3: Return context cancellation from Iter when applicable #62
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
When the context passed to the S3 Iter method is cancelled, the underlying Minio client sometimes immediately closes the results channel and returns without sending the `ctx.Err()` error. Depending on when exactly the context is cancelled, this can result in the `Iter` method not returning any results _and_ not returning an error indicating what happened. Discovered while debugging grafana/mimir#5433 Signed-off-by: Nick Pillitteri <nick.pillitteri@grafana.com>
GiedriusS
approved these changes
Jul 10, 2023
There was a problem hiding this comment.
LGTM, however, could you please also create an upstream issue in minio-go? It seems like this is something that could be solved there. I also saw this in prod where remote object storage was having problems and sometimes the meta fetcher returned 0 metas with no errors. This is probably the cause. Haven't had time to dig into this myself.
Opened an issue, thanks! |
3 tasks
56quarters
added a commit
to grafana/mimir
that referenced
this pull request
Jul 10, 2023
Specifically to pull in thanos-io/objstore#62 which fixes an issue where all store-gateways unloaded all blocks due to a badly timed shutdown. The library has removed the thanos_ prefix from metrics which our dashboards rely on. I've worked around this change by adding the prefix back. Fixes #5433 Signed-off-by: Nick Pillitteri <nick.pillitteri@grafana.com>
56quarters
added a commit
to grafana/mimir
that referenced
this pull request
Jul 11, 2023
Specifically to pull in thanos-io/objstore#62 which fixes an issue where all store-gateways unloaded all blocks due to a badly timed shutdown. The library has removed the thanos_ prefix from metrics which our dashboards rely on. I've worked around this change by adding the prefix back. Fixes #5433 Signed-off-by: Nick Pillitteri <nick.pillitteri@grafana.com>
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.
Changes
This PR returns
ctx.Err()from the S3Itermethod instead ofnilafter exhausting theListObjectschannel.When the context passed to the S3 Iter method is cancelled, the underlying Minio client sometimes immediately closes the results channel and returns without sending the
ctx.Err()error. Depending on when exactly the context is cancelled, this can result in theItermethod not returning any results and not returning an error indicating what happened.Discovered while debugging grafana/mimir#5433
Verification
Tested via Mimir end-to-end test used to reproduce the issue. The S3
Bucketimplementation uses the concrete MinioClientinstead of an interface so writing a unit test would be a challenge without making quite a few changes to theBucket. I can do that if you'd like but it seemed like a lot of churn for a small change.