Store: Add flag ignore-deletion-marks-errors to be able to ignore errors while retrieving deletion marks

[pettersolberg88]

Store: Add flag ignore-deletion-marks-errors to be able to ignore errors while retrieving deletion marks.

Our S3 implementation (Netapp) have intermittent faults that creates time-outs when querying some non-existent objects.
The IgnoreDeletionMarkFilter queries all metrics blocks for the file deletion-mark.json and when store receives an timeout or other error, it crashes. This flag ignores all fetching errors, and makes store not crash.

Fixes errors like this:

{"caller":"grpc.go:164","component":"store","err":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HA07EAKT1YPCMYC6SDHS58S0/deletion-mark.json: Get \"https:<S3-URL>/thanos-metrics/01HA07EAKT1YPCMYC6SDHS58S0/deletion-mark.json\": dial tcp <IP-address>:443: i/o timeout","level":"info","msg":"internal server is shutdown gracefully","service":"gRPC/server","ts":"2023-10-31T07:06:03.987065433Z"}

• I added CHANGELOG entry for this change.
• Change is not relevant to the end user.

Changes

• Add flag ignore-deletion-marks-errors to be able to ignore errors while retrieving deletion marks.
• Log other errors while processing deletion-marks.

Verification

• Tested against against our s3 implementation. Store do not crash.

[yeya24]

The IgnoreDeletionMarkFilter queries all metrics blocks for the file deletion-mark.json and when store receives an timeout or other error, it crashes.

This doesn't sound right. IO timeout shouldn't crash store

[pettersolberg88]

I Agree that IO timeout should not crash thanos store. For me it seems that the error-handling does not handle timeout properly.

Here is a complete log from thanos-store, that is currently crashlooping. We are running two replicas and both are crashlooping running v0.32.5. And the workaround is to delete the whole chunk.

k logs thanos-store-cold-0 -c thanos-store -p {"caller":"factory.go:53","level":"info","msg":"loading bucket configuration","ts":"2024-01-04T05:46:42.136611622Z"} {"caller":"factory.go:35","level":"info","msg":"loading index cache configuration","ts":"2024-01-04T05:46:42.1370288Z"} {"caller":"memcached.go:71","level":"info","msg":"created index cache","ts":"2024-01-04T05:46:42.137888234Z"} {"caller":"options.go:26","level":"info","msg":"disabled TLS, key and cert must be set to enable","protocol":"gRPC","ts":"2024-01-04T05:46:42.13830578Z"} {"caller":"store.go:520","level":"info","msg":"starting store node","ts":"2024-01-04T05:46:42.139639595Z"} {"caller":"intrumentation.go:75","level":"info","msg":"changing probe status","status":"healthy","ts":"2024-01-04T05:46:42.139735228Z"} {"address":"0.0.0.0:10902","caller":"http.go:73","component":"store","level":"info","msg":"listening for requests and metrics","service":"http/server","ts":"2024-01-04T05:46:42.139776183Z"} {"caller":"store.go:418","level":"info","msg":"initializing bucket store","ts":"2024-01-04T05:46:42.139788709Z"} {"address":"[::]:10902","caller":"tls_config.go:274","component":"store","level":"info","msg":"Listening on","service":"http/server","ts":"2024-01-04T05:46:42.139967167Z"} {"address":"[::]:10902","caller":"tls_config.go:277","component":"store","http2":false,"level":"info","msg":"TLS is disabled.","service":"http/server","ts":"2024-01-04T05:46:42.139987835Z"} {"caller":"intrumentation.go:67","level":"warn","msg":"changing probe status","reason":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json: Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <ip>:443: i/o timeout","status":"not-ready","ts":"2024-01-04T07:02:52.44066162Z"} {"caller":"http.go:91","component":"store","err":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json: Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <ip>:443: i/o timeout","level":"info","msg":"internal server is shutting down","service":"http/server","ts":"2024-01-04T07:02:52.440757359Z"} {"caller":"intrumentation.go:56","level":"info","msg":"changing probe status","status":"ready","ts":"2024-01-04T07:02:52.440804592Z"} {"address":"0.0.0.0:10901","caller":"grpc.go:131","component":"store","level":"info","msg":"listening for serving gRPC","service":"gRPC/server","ts":"2024-01-04T07:02:52.440867161Z"} {"caller":"http.go:110","component":"store","err":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json: Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <ip>:443: i/o timeout","level":"info","msg":"internal server is shutdown gracefully","service":"http/server","ts":"2024-01-04T07:02:52.44091739Z"} {"caller":"intrumentation.go:81","level":"info","msg":"changing probe status","reason":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json: Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <ip>:443: i/o timeout","status":"not-healthy","ts":"2024-01-04T07:02:52.440966495Z"} {"caller":"intrumentation.go:67","level":"warn","msg":"changing probe status","reason":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json: Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <ip>:443: i/o timeout","status":"not-ready","ts":"2024-01-04T07:02:52.441015326Z"} {"caller":"grpc.go:138","component":"store","err":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json: Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <ip>:443: i/o timeout","level":"info","msg":"internal server is shutting down","service":"gRPC/server","ts":"2024-01-04T07:02:52.441033605Z"} {"caller":"grpc.go:151","component":"store","level":"info","msg":"gracefully stopping internal server","service":"gRPC/server","ts":"2024-01-04T07:02:52.44105544Z"} {"caller":"grpc.go:164","component":"store","err":"bucket store initial sync: sync block: filter metas: filter blocks marked for deletion: get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json: Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <ip>:443: i/o timeout","level":"info","msg":"internal server is shutdown gracefully","service":"gRPC/server","ts":"2024-01-04T07:02:52.441094163Z"} {"caller":"main.go:161","err":"Get \"https://<S3-URL>/thanos-metrics/01HH319ACH6FYV9Q28010T93YB/deletion-mark.json\": dial tcp <IP-address-to-S3-provider>:443: i/o timeout get file: 01HH319ACH6FYV9Q28010T93YB/deletion-mark.json github.com/thanos-io/thanos/pkg/block/metadata.ReadMarker \t/app/pkg/block/metadata/markers.go:124 github.com/thanos-io/thanos/pkg/block.(*IgnoreDeletionMarkFilter).Filter.func1 \t/app/pkg/block/fetcher.go:859 golang.org/x/sync/errgroup.(*Group).Go.func1 \t/go/pkg/mod/golang.org/x/[email protected]/errgroup/errgroup.go:75 runtime.goexit \t/usr/local/go/src/runtime/asm_amd64.s:1650 filter blocks marked for deletion github.com/thanos-io/thanos/pkg/block.(*IgnoreDeletionMarkFilter).Filter \t/app/pkg/block/fetcher.go:904 github.com/thanos-io/thanos/pkg/block.(*BaseFetcher).fetch \t/app/pkg/block/fetcher.go:475 github.com/thanos-io/thanos/pkg/block.(*MetaFetcher).Fetch \t/app/pkg/block/fetcher.go:514 github.com/thanos-io/thanos/pkg/store.(*BucketStore).SyncBlocks \t/app/pkg/store/bucket.go:556 github.com/thanos-io/thanos/pkg/store.(*BucketStore).InitialSync \t/app/pkg/store/bucket.go:625 main.runStore.func5.1 \t/app/cmd/thanos/store.go:427 github.com/thanos-io/thanos/pkg/runutil.RetryWithLog \t/app/pkg/runutil/runutil.go:97 github.com/thanos-io/thanos/pkg/runutil.Retry \t/app/pkg/runutil/runutil.go:87 main.runStore.func5 \t/app/cmd/thanos/store.go:426 github.com/oklog/run.(*Group).Run.func1 \t/go/pkg/mod/github.com/oklog/[email protected]/group.go:38 runtime.goexit \t/usr/local/go/src/runtime/asm_amd64.s:1650 filter metas github.com/thanos-io/thanos/pkg/block.(*BaseFetcher).fetch \t/app/pkg/block/fetcher.go:476 github.com/thanos-io/thanos/pkg/block.(*MetaFetcher).Fetch \t/app/pkg/block/fetcher.go:514 github.com/thanos-io/thanos/pkg/store.(*BucketStore).SyncBlocks \t/app/pkg/store/bucket.go:556 github.com/thanos-io/thanos/pkg/store.(*BucketStore).InitialSync \t/app/pkg/store/bucket.go:625 main.runStore.func5.1 \t/app/cmd/thanos/store.go:427 github.com/thanos-io/thanos/pkg/runutil.RetryWithLog \t/app/pkg/runutil/runutil.go:97 github.com/thanos-io/thanos/pkg/runutil.Retry \t/app/pkg/runutil/runutil.go:87 main.runStore.func5 \t/app/cmd/thanos/store.go:426 github.com/oklog/run.(*Group).Run.func1 \t/go/pkg/mod/github.com/oklog/[email protected]/group.go:38 runtime.goexit \t/usr/local/go/src/runtime/asm_amd64.s:1650 sync block github.com/thanos-io/thanos/pkg/store.(*BucketStore).InitialSync \t/app/pkg/store/bucket.go:626 main.runStore.func5.1 \t/app/cmd/thanos/store.go:427 github.com/thanos-io/thanos/pkg/runutil.RetryWithLog \t/app/pkg/runutil/runutil.go:97 github.com/thanos-io/thanos/pkg/runutil.Retry \t/app/pkg/runutil/runutil.go:87 main.runStore.func5 \t/app/cmd/thanos/store.go:426 github.com/oklog/run.(*Group).Run.func1 \t/go/pkg/mod/github.com/oklog/[email protected]/group.go:38 runtime.goexit \t/usr/local/go/src/runtime/asm_amd64.s:1650 bucket store initial sync main.runStore.func5 \t/app/cmd/thanos/store.go:432 github.com/oklog/run.(*Group).Run.func1 \t/go/pkg/mod/github.com/oklog/[email protected]/group.go:38 runtime.goexit \t/usr/local/go/src/runtime/asm_amd64.s:1650 store command failed main.main \t/app/cmd/thanos/main.go:161 runtime.main \t/usr/local/go/src/runtime/proc.go:267 runtime.goexit \t/usr/local/go/src/runtime/asm_amd64.s:1650","level":"error","ts":"2024-01-04T07:02:52.441344799Z"}