[jira] [Commented] (CASSANDRA-8639) Can OOM on CL replay with dense mutations

["Ariel Weisberg (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-8639?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15025606#comment-15025606 ] 

Ariel Weisberg commented on CASSANDRA-8639:
-------------------------------------------

Proposed change. Adding a limit on the number of bytes was easy. Writing a unit test was not so easy. I refactored CommitLogReplayer so you can spy on the mutations it initiates. Suggestions for a better way to test this new behavior are welcome.

|[2.1 code|https://github.com/apache/cassandra/compare/cassandra-2.1...aweisberg:CASSANDRA-8639-2.1?expand=1]|[utests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-CASSANDRA-8639-2.1-testall/]|[dtests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-CASSANDRA-8639-2.1-dtest/]|
|[2.2 code|https://github.com/apache/cassandra/compare/cassandra-2.2...aweisberg:CASSANDRA-8639-2.2?expand=1]|[utests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-CASSANDRA-8639-2.2-testall/]|[dtests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-CASSANDRA-8639-2.2-dtest/]|
|[3.0 code|https://github.com/apache/cassandra/compare/cassandra-3.0...aweisberg:CASSANDRA-8639-3.0?expand=1]|[utests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-CASSANDRA-8639-3.0-testall/]|[dtests|https://cassci.datastax.com/view/Dev/view/aweisberg/job/aweisberg-CASSANDRA-8639-3.0-dtest/]|

> Can OOM on CL replay with dense mutations
> -----------------------------------------
>
>                 Key: CASSANDRA-8639
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-8639
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: T Jake Luciani
>            Assignee: Ariel Weisberg
>            Priority: Minor
>             Fix For: 2.1.x
>
>
> If you write dense mutations with many clustering keys, the replay of the CL can quickly overwhelm a node on startup.  This looks to be caused by the fact we only ensure there are 1000 mutations in flight at a time. but those mutations could have thousands of cells in them.
> A better approach would be to limit the CL replay to the amount of memory in flight using cell.unsharedHeapSize()



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)