[
https://issues.apache.org/jira/browse/CASSANDRA-20039?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17894368#comment-17894368
]
Brandon Williams commented on CASSANDRA-20039:
----------------------------------------------
I suspect you have corruption somewhere which is throwing off the compression
and that is being exposed by the bootstrapping node when streaming it.
Unfortunately it's not easy to tell where it is from the log.
> Memory leak in Bootstraping
> ---------------------------
>
> Key: CASSANDRA-20039
> URL: https://issues.apache.org/jira/browse/CASSANDRA-20039
> Project: Cassandra
> Issue Type: Bug
> Reporter: Aleksandr Batyrshin
> Priority: Normal
> Attachments: 2024-10-31_01-23-24-resized.png,
> system.log.2024-10-28.0.zip
>
>
> We recently updated Cassandra to 4.1.7 from 4.0.3 and noticed that bootstrap
> started to consume a lot of memory.
> I have a feeling that there is a memory leak somewhere, because even 32GB of
> heap is not enough and the node becomes unresponsive (long pauses on the GC).
> !2024-10-31_01-23-24-resized.png!
> A simple analysis (jmap -histo:live) shows that we have a constantly growing
> org.apache.cassandra.io.compress.CompressionMetadata$Chunk
> If required, I am ready to provide additional information
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
