[ https://issues.apache.org/jira/browse/KAFKA-15609?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17782148#comment-17782148 ]
Divij Vaidya commented on KAFKA-15609: -------------------------------------- Hey folks I did some more digging on this and I am not convinced that this is not a bug but I don't have proof right now. My understanding is as follow: 1. MMap will read/write from page cache in the OS. 2. MMap flushed data to page cache "eventually" after writes. Note that MappedByteBuffer.force() method forces a flush of MMap to page cache. 3. In Kafka, we flush indexes leading to MappedByteBuffer.force() on segment roll, but this is performed asynchronously in a separate thread. Now consider the situation: 1. Data is written to segment and index. 2. Segment rolls and schedules a flush. 3. Since segment is closed, RSM will consider it for copying to remote. 4. RSM will read from page cache which is guaranteed to give correct (even non-flushed / dirty) data for segment but it may give out of date data for indexes. Result: Indexes uploaded to remote tier will not contain all data that is expected in it. I will keep this ticket closed until I have some reproducer but I would solicit your thoughts on this. Proposal: To fix this, we should call AbstractIndex.flush() in RLM before passing the index to RSM for upload. > Corrupted index uploaded to remote tier > --------------------------------------- > > Key: KAFKA-15609 > URL: https://issues.apache.org/jira/browse/KAFKA-15609 > Project: Kafka > Issue Type: Bug > Components: Tiered-Storage > Affects Versions: 3.6.0 > Reporter: Divij Vaidya > Priority: Minor > > While testing Tiered Storage, we have observed corrupt indexes being present > in remote tier. One such situation is covered here at > https://issues.apache.org/jira/browse/KAFKA-15401. This Jira presents another > such possible case of corruption. > Potential cause of index corruption: > We want to ensure that the file we are passing to RSM plugin contains all the > data which is present in MemoryByteBuffer i.e. we should have flushed the > MemoryByteBuffer to the file using force(). In Kafka, when we close a > segment, indexes are flushed asynchronously [1]. Hence, it might be possible > that when we are passing the file to RSM, the file doesn't contain flushed > data. Hence, we may end up uploading indexes which haven't been flushed yet. > Ideally, the contract should enforce that we force flush the content of > MemoryByteBuffer before we give the file for RSM. This will ensure that > indexes are not corrupted/incomplete. > [1] > [https://github.com/apache/kafka/blob/4150595b0a2e0f45f2827cebc60bcb6f6558745d/core/src/main/scala/kafka/log/UnifiedLog.scala#L1613] > -- This message was sent by Atlassian Jira (v8.20.10#820010)