Fix Snapshots Recording Incorrect Max. Segment Counts #74291
Merged
Conversation
This file contains hidden or 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
If sequence numbers are equal across snapshots (and thus no files get written to the repository) the segment count in the index commit is not reliable as it may have changed due to background merges. => fixed by always using the segment count determined from the file names in the snapshot instead closes elastic#74249
original-brownbear
added
>bug
:Distributed Coordination/Snapshot/Restore
elasticmachine
added
the
Team:Distributed (Obsolete)
|
Pinging @elastic/es-distributed (Team:Distributed) |
|
Could we set |
Done, so much cuter :) Thanks! |
DaveCTurner
reviewed
Jun 18, 2021
| final int rounds = scaledRandomIntBetween(50, 300); | ||
| for (int i = 0; i < rounds; ++i) { | ||
| final int numDocs = scaledRandomIntBetween(100, 1000); | ||
| BulkRequestBuilder request = client().prepareBulk(); |
There was a problem hiding this comment.
I think it'd be better to refresh on each bulk to force a segment to be written:
Suggested change
| BulkRequestBuilder request = client().prepareBulk(); | |
| BulkRequestBuilder request = client().prepareBulk().setRefreshPolicy(WriteRequest.RefreshPolicy.IMMEDIATE); |
I suspect if we did that we wouldn't need so many rounds.
There was a problem hiding this comment.
++ right done in d629c74
DaveCTurner
approved these changes
Jun 18, 2021
| .prepareUpdateSettings(indexName) | ||
| .setSettings( | ||
| Settings.builder() | ||
| .put(MergePolicyConfig.INDEX_MERGE_POLICY_SEGMENTS_PER_TIER_SETTING.getKey(), "2") |
There was a problem hiding this comment.
👍 we know we did ≥3 rounds, each of which made ≥1 segment, and they're all pretty small (≤2MB) so this guarantees a merge IIUC.
There was a problem hiding this comment.
++ that was assumption here. Also ran a few hundred rounds of this to verify and it seems to hold 🤞
|
Thanks David! |
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Jun 18, 2021
If sequence numbers are equal across snapshots (and thus no files get written to the repository) the segment count in the index commit is not reliable as it may have changed due to background merges. => fixed by always using the segment count determined from the file names in the snapshot instead closes elastic#74249
original-brownbear
added a commit
to original-brownbear/elasticsearch
that referenced
this pull request
Jun 18, 2021
If sequence numbers are equal across snapshots (and thus no files get written to the repository) the segment count in the index commit is not reliable as it may have changed due to background merges. => fixed by always using the segment count determined from the file names in the snapshot instead closes elastic#74249
original-brownbear
added a commit
that referenced
this pull request
Jun 18, 2021
If sequence numbers are equal across snapshots (and thus no files get written to the repository) the segment count in the index commit is not reliable as it may have changed due to background merges. => fixed by always using the segment count determined from the file names in the snapshot instead closes #74249
original-brownbear
added a commit
that referenced
this pull request
Jun 18, 2021
If sequence numbers are equal across snapshots (and thus no files get written to the repository) the segment count in the index commit is not reliable as it may have changed due to background merges. => fixed by always using the segment count determined from the file names in the snapshot instead closes #74249
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.
If sequence numbers are equal across snapshots (and thus no files get written to the repository)
the segment count in the index commit is not reliable as it may have changed due to background merges.
=> fixed by always using the segment count determined from the file names in the snapshot instead
closes #74249
I'm not the biggest fan of my test here to be honest, but I couldn't find a way to have more fine grained control over background merges so I went with this rather brute force approach (alternative test suggestions very welcome :)). The test does however reproduce the problem reliable in 100% of runs for me.