Initialize sequence numbers on a shrunken index #25321
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
Bringing together shards in a shrunken index means that we need to address the start of history for the shrunken index. The problem here is that sequence numbers before the maximum of the maximum sequence numbers on the source shards can collide in the target shards in the shrunken index. To address this, we set the maximum sequence number and the local checkpoint on the target shards to this maximum of the maximum sequence numbers. This enables correct document-level semantics for documents indexed before the shrink, and history on the shrunken index will effectively start from here.
64 tasks
bleskes
approved these changes
Jun 21, 2017
| .collect(Collectors.toList()).toArray(new Directory[shards.size()])); | ||
|
|
||
| final Directory[] sources = | ||
| shards.stream().map(LocalShardSnapshot::getSnapshotDirectory).collect(Collectors.toList()).toArray(new Directory[0]); |
Contributor
There was a problem hiding this comment.
Nit - there is a stream::toArray(Directory[]::new))
| writer.setLiveCommitData(() -> { | ||
| final HashMap<String, String> liveCommitData = new HashMap<>(2); | ||
| liveCommitData.put(SequenceNumbers.MAX_SEQ_NO, Long.toString(maxSeqNo)); | ||
| liveCommitData.put(SequenceNumbers.LOCAL_CHECKPOINT_KEY, Long.toString(maxSeqNo)); |
Contributor
There was a problem hiding this comment.
Is the plan to do MAX_UNSAFE_AUTO_ID_TIMESTAMP_COMMIT_ID as a follow up?
Member
Author
There was a problem hiding this comment.
Yes, I edited the plan on #10708 to separate these out into separate line items, I do not like mixing things. I do forgive you for not seeing this. 😛
| .put("index.version.created", version) | ||
| ).get(); | ||
| for (int i = 0; i < 20; i++) { | ||
| final int docs = randomIntBetween(1, 128); |
Contributor
There was a problem hiding this comment.
do we want to test with 0 docs too?
clintongormley
added
:Distributed Indexing/Engine
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 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.
Bringing together shards in a shrunken index means that we need to address the start of history for the shrunken index. The problem here is that sequence numbers before the maximum of the maximum sequence numbers on the source shards can collide in the target shards in the shrunken index. To address this, we set the maximum sequence number and the local checkpoint on the target shards to this maximum of the maximum sequence numbers. This enables correct document-level semantics for documents indexed before the shrink, and history on the shrunken index will effectively start from here.
Relates #10708