You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@wicket.apache.org by "Ondra Žižka (Created JIRA)" <ji...@apache.org> on 2012/03/27 17:26:27 UTC
[jira] [Created] (WICKET-4478) DiskDataStore to use multi-level
directory structure to avoid slowness when thousands of sessions are
active.
DiskDataStore to use multi-level directory structure to avoid slowness when thousands of sessions are active.
-------------------------------------------------------------------------------------------------------------
Key: WICKET-4478
URL: https://issues.apache.org/jira/browse/WICKET-4478
Project: Wicket
Issue Type: Improvement
Components: wicket
Affects Versions: 1.5.5
Environment: ext3, NTFS, and common FSs in general.
Reporter: Ondra Žižka
If the number of sessions opened rises to thousands, DiskDataStore, which is default, doesn't scale, because it creates one dir for each session, and most filesystems don't perform well with large number of records in a dir. More, there's a hard limit e.g. for ext3 it's 31998.
So I suggest to make it multi-level - simply, instead of "tmp/abcdef" it would be "tmp/a/b/cdef", where number of levels added would be configurable.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Updated] (WICKET-4478) DiskDataStore to use multi-level
directory structure to avoid slowness when thousands of sessions are
active.
Posted by "Ondra Žižka (Updated JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/WICKET-4478?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ondra Žižka updated WICKET-4478:
--------------------------------
Attachment: fix-WICKET-4478.patch
Something like this.
It's a 2-level dir, using first char.
> DiskDataStore to use multi-level directory structure to avoid slowness when thousands of sessions are active.
> -------------------------------------------------------------------------------------------------------------
>
> Key: WICKET-4478
> URL: https://issues.apache.org/jira/browse/WICKET-4478
> Project: Wicket
> Issue Type: Improvement
> Components: wicket
> Affects Versions: 1.5.5
> Environment: ext3, NTFS, and common FSs in general.
> Reporter: Ondra Žižka
> Attachments: fix-WICKET-4478.patch
>
>
> If the number of sessions opened rises to thousands, DiskDataStore, which is default, doesn't scale, because it creates one dir for each session, and most filesystems don't perform well with large number of records in a dir. More, there's a hard limit e.g. for ext3 it's 31998.
> So I suggest to make it multi-level - simply, instead of "tmp/abcdef" it would be "tmp/a/b/cdef", where number of levels added would be configurable.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Resolved] (WICKET-4478) DiskDataStore to use multi-level
directory structure to avoid slowness when thousands of sessions are
active.
Posted by "Martin Grigorov (Resolved) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/WICKET-4478?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Martin Grigorov resolved WICKET-4478.
-------------------------------------
Resolution: Fixed
Fix Version/s: 6.0.0-RC1
1.5.6
Improved by re-applying the patch from WICKET-1837.
It uses the hashcode of the sessionId string to make it better distributed.
> DiskDataStore to use multi-level directory structure to avoid slowness when thousands of sessions are active.
> -------------------------------------------------------------------------------------------------------------
>
> Key: WICKET-4478
> URL: https://issues.apache.org/jira/browse/WICKET-4478
> Project: Wicket
> Issue Type: Improvement
> Components: wicket
> Affects Versions: 1.5.5
> Environment: ext3, NTFS, and common FSs in general.
> Reporter: Ondra Žižka
> Assignee: Martin Grigorov
> Fix For: 1.5.6, 6.0.0-RC1
>
> Attachments: fix-WICKET-4478.patch
>
>
> If the number of sessions opened rises to thousands, DiskDataStore, which is default, doesn't scale, because it creates one dir for each session, and most filesystems don't perform well with large number of records in a dir. More, there's a hard limit e.g. for ext3 it's 31998.
> So I suggest to make it multi-level - simply, instead of "tmp/abcdef" it would be "tmp/a/b/cdef", where number of levels added would be configurable.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
[jira] [Assigned] (WICKET-4478) DiskDataStore to use multi-level
directory structure to avoid slowness when thousands of sessions are
active.
Posted by "Martin Grigorov (Assigned) (JIRA)" <ji...@apache.org>.
[ https://issues.apache.org/jira/browse/WICKET-4478?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Martin Grigorov reassigned WICKET-4478:
---------------------------------------
Assignee: Martin Grigorov
> DiskDataStore to use multi-level directory structure to avoid slowness when thousands of sessions are active.
> -------------------------------------------------------------------------------------------------------------
>
> Key: WICKET-4478
> URL: https://issues.apache.org/jira/browse/WICKET-4478
> Project: Wicket
> Issue Type: Improvement
> Components: wicket
> Affects Versions: 1.5.5
> Environment: ext3, NTFS, and common FSs in general.
> Reporter: Ondra Žižka
> Assignee: Martin Grigorov
> Attachments: fix-WICKET-4478.patch
>
>
> If the number of sessions opened rises to thousands, DiskDataStore, which is default, doesn't scale, because it creates one dir for each session, and most filesystems don't perform well with large number of records in a dir. More, there's a hard limit e.g. for ext3 it's 31998.
> So I suggest to make it multi-level - simply, instead of "tmp/abcdef" it would be "tmp/a/b/cdef", where number of levels added would be configurable.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators: https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira