Allow leaving corrupted indexes as readable after header repair#3801
Merged
ScottDugas merged 19 commits intoFoundationDB:mainfrom Mar 2, 2026
Merged
Allow leaving corrupted indexes as readable after header repair#3801ScottDugas merged 19 commits intoFoundationDB:mainfrom
ScottDugas merged 19 commits intoFoundationDB:mainfrom
Conversation
In the case of a broader corruption, where records have been corrupted along with the store header, it may be beneficial to look at the indexes involved while looking at the store. This may be useful if you want to resurrect some information from the indexes if any exists.
ohadzeliger
requested changes
Dec 19, 2025
…tBase As noted in the PR `withStore` was potentially brittle because of the `recordStore` field. I could have changed that to a consumer but still set the field, but that felt just as bad, so I changed it to extend FDBRecordStoreConcurrentTestBase.
ohadzeliger
approved these changes
Jan 2, 2026
Conflict on PreventCommitCheck because that was brought into main on a different PR. Conflict on FDBRecordStoreRepairHeaderTest because I added a method close to a method that was renamed. I changed the new test method because the rename made it clear that createOriginalRecords didn't create anything, and thus the call was unnecessary
Also, I had to fix a couple tests that dependended on the recordStore field.
Primarily by adding helper methods
This makes it easier to cover the version that doesn't take a parameter.
I'm not sure whether we should make the assertion stronger when we expect there to be data.
ScottDugas
commented
Feb 11, 2026
| @API(API.Status.DEPRECATED) | ||
| public CompletableFuture<NonnullPair<Boolean, FDBRecordStore>> repairMissingHeader(final int userVersion, FormatVersion minimumPossibleFormatVersion) { | ||
| return repairMissingHeader(userVersion, minimumPossibleFormatVersion, null); | ||
| } |
Collaborator
Author
There was a problem hiding this comment.
I looked into this test gap:
I think I'm seeing a false test gap. https://fdb.teamscale.io/activity/merge-requests/foundationdb-fdb-record-layer/FoundationDB%2Ffdb-record-layer%2F3801 for commit 049a768 I checked the upload details, and downloaded the coverage, and all the repairMissingHeader methods have full method coverage: https://fdb.teamscale.io/project/project?name=foundationdb-fdb-record-layer&action=showExternalAnalysisUploads -- https://fdb.teamscale.io/api/projects/foundationdb-fdb-record-layer/external-analysis/reports/FORK_MR%2F3801%2FScottDugas%2Frepair-store-without-disable%3A1770831010000/0
Oh, I see, in the coverage provided it is at 5811:
<method name="repairMissingHeader" desc="(ILcom/apple/foundationdb/record/provider/foundationdb/FormatVersion;)Ljava/util/concurrent/CompletableFuture;" line="5811">
<counter type="INSTRUCTION" missed="0" covered="6"/>
<counter type="LINE" missed="0" covered="1"/>
<counter type="COMPLEXITY" missed="0" covered="1"/>
<counter type="METHOD" missed="0" covered="1"/>
</method>
And if I look at the html from jacoco, you can see that the line numbers are different:
ohadzeliger
requested changes
Feb 16, 2026
Contributor
There was a problem hiding this comment.
Seems like we're already doing this?
Collaborator
Author
There was a problem hiding this comment.
Partially, I think the fact that IndexScrubbing only works for value indexes is still valuable.
| } | ||
|
|
||
| enum DisableIndexBehavior { | ||
| LeaveCorrupted { |
Instead of having a @nullable reason, this means using a method, which makes it a lot clearer what you are intending.
Most notably, adding a missing join() in the test code
ScottDugas
force-pushed
the
repair-store-without-disable
branch
from
1376463 to
4ad934c
Compare
ohadzeliger
approved these changes
Feb 27, 2026
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
In the case of a broader corruption, where records have been corrupted along with the store header, it may be beneficial to look at the indexes involved while looking at the store. This may be useful if you want to resurrect some information from the indexes if any exists.