You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by "Brandon Williams (JIRA)" <ji...@apache.org> on 2011/09/09 21:48:09 UTC

[jira] [Created] (CASSANDRA-3171) AbstractCompactionIterable uses a 1MB buffer for every sstable

AbstractCompactionIterable uses a 1MB buffer for every sstable
--------------------------------------------------------------

                 Key: CASSANDRA-3171
                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3171
             Project: Cassandra
          Issue Type: Bug
          Components: Core
    Affects Versions: 1.0.0
            Reporter: Brandon Williams
            Assignee: Brandon Williams
             Fix For: 1.0.0


This causes an OOM for any compaction task that needs to open all the sstables when you have a lot of them (repair, major, etc)

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] [Updated] (CASSANDRA-3171) AbstractCompactionIterable uses a 1MB buffer for every sstable

Posted by "Brandon Williams (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3171?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Brandon Williams updated CASSANDRA-3171:
----------------------------------------

    Attachment: 3171.txt

Patch to drop buffer size to 64k, as is the RandomAccessReader default.

> AbstractCompactionIterable uses a 1MB buffer for every sstable
> --------------------------------------------------------------
>
>                 Key: CASSANDRA-3171
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3171
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.0.0
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>             Fix For: 1.0.0
>
>         Attachments: 3171.txt
>
>
> This causes an OOM for any compaction task that needs to open all the sstables when you have a lot of them (repair, major, etc)

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] [Commented] (CASSANDRA-3171) AbstractCompactionIterable uses a 1MB buffer for every sstable

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3171?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101524#comment-13101524 ] 

Jonathan Ellis commented on CASSANDRA-3171:
-------------------------------------------

Is there anywhere we actually need non-default buffer sizes?  I think we should probably rip out that constructor parameter (on both RAR and SW) to force everyone to be sane whether they like it or not.

See also: the 10MB buffer in streaming and the 8MB one in IndexWriter.

> AbstractCompactionIterable uses a 1MB buffer for every sstable
> --------------------------------------------------------------
>
>                 Key: CASSANDRA-3171
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3171
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.0.0
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>             Fix For: 1.0.0
>
>         Attachments: 3171.txt
>
>
> This causes an OOM for any compaction task that needs to open all the sstables when you have a lot of them (repair, major, etc)

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] [Updated] (CASSANDRA-3171) AbstractCompactionIterable uses a 1MB buffer for every sstable

Posted by "Brandon Williams (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3171?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Brandon Williams updated CASSANDRA-3171:
----------------------------------------

    Reviewer: jbellis

> AbstractCompactionIterable uses a 1MB buffer for every sstable
> --------------------------------------------------------------
>
>                 Key: CASSANDRA-3171
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3171
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.0.0
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>             Fix For: 1.0.0
>
>         Attachments: 3171.txt
>
>
> This causes an OOM for any compaction task that needs to open all the sstables when you have a lot of them (repair, major, etc)

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] [Commented] (CASSANDRA-3171) AbstractCompactionIterable uses a 1MB buffer for every sstable

Posted by "Brandon Williams (JIRA)" <ji...@apache.org>.
    [ https://issues.apache.org/jira/browse/CASSANDRA-3171?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13101570#comment-13101570 ] 

Brandon Williams commented on CASSANDRA-3171:
---------------------------------------------

+1

> AbstractCompactionIterable uses a 1MB buffer for every sstable
> --------------------------------------------------------------
>
>                 Key: CASSANDRA-3171
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3171
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.0.0
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>             Fix For: 1.0.0
>
>         Attachments: 3171-v2.txt, 3171.txt
>
>
> This causes an OOM for any compaction task that needs to open all the sstables when you have a lot of them (repair, major, etc)

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] [Resolved] (CASSANDRA-3171) AbstractCompactionIterable uses a 1MB buffer for every sstable

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3171?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis resolved CASSANDRA-3171.
---------------------------------------

    Resolution: Fixed

committed

> AbstractCompactionIterable uses a 1MB buffer for every sstable
> --------------------------------------------------------------
>
>                 Key: CASSANDRA-3171
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3171
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.0.0
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>             Fix For: 1.0.0
>
>         Attachments: 3171-v2.txt, 3171.txt
>
>
> This causes an OOM for any compaction task that needs to open all the sstables when you have a lot of them (repair, major, etc)

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

[jira] [Updated] (CASSANDRA-3171) AbstractCompactionIterable uses a 1MB buffer for every sstable

Posted by "Jonathan Ellis (JIRA)" <ji...@apache.org>.
     [ https://issues.apache.org/jira/browse/CASSANDRA-3171?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Jonathan Ellis updated CASSANDRA-3171:
--------------------------------------

    Attachment: 3171-v2.txt

v2 removes unnecessary bufferSize parameters across the call heirarchy.  Main exception is CRAR, to which I added

{code}
// TODO refactor this to separate concept of "buffer to avoid lots of read() syscalls" and "compression buffer"
{code}

> AbstractCompactionIterable uses a 1MB buffer for every sstable
> --------------------------------------------------------------
>
>                 Key: CASSANDRA-3171
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-3171
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Core
>    Affects Versions: 1.0.0
>            Reporter: Brandon Williams
>            Assignee: Brandon Williams
>             Fix For: 1.0.0
>
>         Attachments: 3171-v2.txt, 3171.txt
>
>
> This causes an OOM for any compaction task that needs to open all the sstables when you have a lot of them (repair, major, etc)

--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira