[jira] [Resolved] (DIRSERVER-1723) Potential LRUCache exhaustion with searches

["Emmanuel Lecharny (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/DIRSERVER-1723?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Emmanuel Lecharny resolved DIRSERVER-1723.
------------------------------------------

    Resolution: Fixed

Solved in the trunk, still an issue in the txns-branch
                
> Potential LRUCache exhaustion with searches
> -------------------------------------------
>
>                 Key: DIRSERVER-1723
>                 URL: https://issues.apache.org/jira/browse/DIRSERVER-1723
>             Project: Directory ApacheDS
>          Issue Type: Bug
>    Affects Versions: 2.0.0-M6
>            Reporter: Emmanuel Lecharny
>            Priority: Blocker
>             Fix For: 2.0.0-M8
>
>
> There are many case where the server will freeze if users don't correctly close the cursors they get when doing a search :
> - in JNDI, but forgotten to close a NamingEnumeration (*very* frequent use case !)
> - when many users create PagedSearches that last for a very long period of time...
> - when a client brutally disconnect (no alert will be sent to the server, letting the cursor pending until the timeout is reached)
> There is no way we can solve those issues by increasing the LRUCache size.
> The only possible solution would be to allow some search to be proceeded outside of any read transaction. That would also imply we deal with ID which have been invalidated due to a modification. In fact, a Search should allow phantom reads.
> In any case, for internal searches, we should keep a transaction support for searches.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira