Fix the reader skips compacted data which original ledger been removed - #12522
Merged
merlimat merged 4 commits intoNov 2, 2021
Conversation
The compactor update the compaction cursor(mark delete) first and then update the `compactionHorizon` of the compacted topic. During the compaction cursor move forward, the original ledger will be removed if no other durable cursors. At the same time, if the reader is reading data from the original ledger, the reader will skip the data while the original ledger been removed, details to see apache#6787. So the reader might skip the compacted data since the `compactionHorizon` have not updated yet. The approach is: 1. Update the `compactionHorizon` before the compaction cursor move forward, so that the reader will not skip the original data before `compactionHorizon` updated. If the broker crashes before the new compacted Ledger ID been persistent, after the topic been loaded, the compaction can be trigger again and will not loss any data, but we will have an orphan ledger cannot be delete in the BookKeeper cluster. 2. Remove the previous compacted Ledger after the compaction cursor move forward, make sure the new compacted Ledger ID been persistent, Otherwise, we might lost compacted ledger if broker crashes.
codelipenghui
added
type/bug
merlimat
approved these changes
Oct 28, 2021
codelipenghui
deleted the
penghui/fix-skip-compaction-data-while-remove-original-ledger
branch
November 3, 2021 11:31
hangc0276
pushed a commit
that referenced
this pull request
Nov 4, 2021
#12522) * Fix the reader skips compacted data which original ledger been removed The compactor update the compaction cursor(mark delete) first and then update the `compactionHorizon` of the compacted topic. During the compaction cursor move forward, the original ledger will be removed if no other durable cursors. At the same time, if the reader is reading data from the original ledger, the reader will skip the data while the original ledger been removed, details to see #6787. So the reader might skip the compacted data since the `compactionHorizon` have not updated yet. The approach is: 1. Update the `compactionHorizon` before the compaction cursor move forward, so that the reader will not skip the original data before `compactionHorizon` updated. If the broker crashes before the new compacted Ledger ID been persistent, after the topic been loaded, the compaction can be trigger again and will not loss any data, but we will have an orphan ledger cannot be delete in the BookKeeper cluster. 2. Remove the previous compacted Ledger after the compaction cursor move forward, make sure the new compacted Ledger ID been persistent, Otherwise, we might lost compacted ledger if broker crashes. * Fix checkstyle * Fix tests. * Fix test (cherry picked from commit 74dd9b9)
eolivelli
pushed a commit
to eolivelli/pulsar
that referenced
this pull request
Nov 29, 2021
apache#12522) * Fix the reader skips compacted data which original ledger been removed The compactor update the compaction cursor(mark delete) first and then update the `compactionHorizon` of the compacted topic. During the compaction cursor move forward, the original ledger will be removed if no other durable cursors. At the same time, if the reader is reading data from the original ledger, the reader will skip the data while the original ledger been removed, details to see apache#6787. So the reader might skip the compacted data since the `compactionHorizon` have not updated yet. The approach is: 1. Update the `compactionHorizon` before the compaction cursor move forward, so that the reader will not skip the original data before `compactionHorizon` updated. If the broker crashes before the new compacted Ledger ID been persistent, after the topic been loaded, the compaction can be trigger again and will not loss any data, but we will have an orphan ledger cannot be delete in the BookKeeper cluster. 2. Remove the previous compacted Ledger after the compaction cursor move forward, make sure the new compacted Ledger ID been persistent, Otherwise, we might lost compacted ledger if broker crashes. * Fix checkstyle * Fix tests. * Fix test
1 task
eolivelli
pushed a commit
that referenced
this pull request
Dec 13, 2021
#12522) * Fix the reader skips compacted data which original ledger been removed The compactor update the compaction cursor(mark delete) first and then update the `compactionHorizon` of the compacted topic. During the compaction cursor move forward, the original ledger will be removed if no other durable cursors. At the same time, if the reader is reading data from the original ledger, the reader will skip the data while the original ledger been removed, details to see #6787. So the reader might skip the compacted data since the `compactionHorizon` have not updated yet. The approach is: 1. Update the `compactionHorizon` before the compaction cursor move forward, so that the reader will not skip the original data before `compactionHorizon` updated. If the broker crashes before the new compacted Ledger ID been persistent, after the topic been loaded, the compaction can be trigger again and will not loss any data, but we will have an orphan ledger cannot be delete in the BookKeeper cluster. 2. Remove the previous compacted Ledger after the compaction cursor move forward, make sure the new compacted Ledger ID been persistent, Otherwise, we might lost compacted ledger if broker crashes. * Fix checkstyle * Fix tests. * Fix test (cherry picked from commit 74dd9b9)
eolivelli
pushed a commit
to eolivelli/pulsar
that referenced
this pull request
Feb 25, 2022
apache#12522) * Fix the reader skips compacted data which original ledger been removed The compactor update the compaction cursor(mark delete) first and then update the `compactionHorizon` of the compacted topic. During the compaction cursor move forward, the original ledger will be removed if no other durable cursors. At the same time, if the reader is reading data from the original ledger, the reader will skip the data while the original ledger been removed, details to see apache#6787. So the reader might skip the compacted data since the `compactionHorizon` have not updated yet. The approach is: 1. Update the `compactionHorizon` before the compaction cursor move forward, so that the reader will not skip the original data before `compactionHorizon` updated. If the broker crashes before the new compacted Ledger ID been persistent, after the topic been loaded, the compaction can be trigger again and will not loss any data, but we will have an orphan ledger cannot be delete in the BookKeeper cluster. 2. Remove the previous compacted Ledger after the compaction cursor move forward, make sure the new compacted Ledger ID been persistent, Otherwise, we might lost compacted ledger if broker crashes. * Fix checkstyle * Fix tests. * Fix test (cherry picked from commit 74dd9b9) (cherry picked from commit df68493)
codelipenghui
restored the
penghui/fix-skip-compaction-data-while-remove-original-ledger
branch
May 17, 2022 01:23
codelipenghui
deleted the
penghui/fix-skip-compaction-data-while-remove-original-ledger
branch
May 17, 2022 01:29
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
7 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.
The compactor update the compaction cursor(mark delete) first and then update the
compactionHorizonof the compacted topic. During the compaction cursor move forward, the original ledger will be removed if no other durable cursors. At the same time, if the reader is reading data from the original ledger, the reader will skip the data while the original ledger been removed, details to see #6787. So the reader might skip the compacted data since thecompactionHorizonhave not updated yet.The approach is:
compactionHorizonbefore the compaction cursor move forward,so that the reader will not skip the original data before
compactionHorizonupdated.If the broker crashes before the new compacted Ledger ID been persistent,
after the topic been loaded, the compaction can be trigger again and will not loss any data,
but we will have an orphan ledger cannot be delete in the BookKeeper cluster.
doc-not-needed