[
https://issues.apache.org/jira/browse/JCR-929?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12499711
]
Dominique Pfister commented on JCR-929:
---------------------------------------
Hi Ian,
thank you for your stack traces and all your work! In your second thread dump,
I'd say that the first one (Thread http-8080-Processor25) still holds the
AbstractJournal's RWLock (acquired in SharedItemStateManager$Update.begin) and
therefore the VM's state is similar to the first thread dump you provided: one
thread holds the AbstractJournal's RWLock and will start an item update (1),
while other threads interoperate with the LockManager and therefore lock that
one (2). When the item update (1) triggers a synchronization on the journal
(because another instance made some changes) it might encounter a lock
operation and will try to inform the LockManager about this event. Because of
all other threads in (2) this will cause the deadlock.
IMO, to solve this problem, LockManager operations will have to adopt the same
pattern as SharedItemStateManager updates already do: lock-and-sync the journal
when the operation starts, unlock at the end of it.
Kind regards
Dominique
> Under Heavy load in a Cluster HTTP Threads Block and stall requests
> -------------------------------------------------------------------
>
> Key: JCR-929
> URL: https://issues.apache.org/jira/browse/JCR-929
> Project: Jackrabbit
> Issue Type: Bug
> Components: core
> Affects Versions: 1.3
> Environment: 2 Node Cluster, OSX, JDK 1.5 with DatabaseJournal,
> DatabasePersistanceManager, all content in DB, using WebDAV to load
> Reporter: Ian Boston
> Assignee: Dominique Pfister
> Attachments: catalina.out.node1.txt, catalina.out.node2.txt
>
>
> Under Heavy load created by mounting both nodes in the cluster in OSX Finder
> and then uploading large numebers of files to each node at the same time ( a
> few 1000), eventually one of the nodes stops responding and the Finder mount
> timesout and disconnects.
> Once that happens that node becomes unusable.
> More mount attempts will prompt for a password indicating HTTP is still
> running, but will timeout once the connection is authenticated.
> Access by the Web Browser will prompt for a password, conenct and provide a
> once only listing of any collection in the workspace. If you try to refresh
> that collection, the HTTP request hangs forever.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
