Kevin,
Thanks for the response. We weren't experiencing deadlocks either, just a significant number of threads waiting on a lock for a fairly long time. Yes we are on WebLogic, and although it comes with a version of JPA we are not using the WebLogic packaged version, but instead OpenJPA 1.2.1 I may try and get another thread dump, I believe this problem was much reduced by upgrading the JDK (due to this issue: http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6461827) Simon Kevin Sutter wrote: > > Hi Simon, > > On Tue, Jun 2, 2009 at 3:33 PM, Simon Droscher < > simon.drosc...@elasticpath.com> wrote: > >> >> Kevin Sutter wrote: >> > >> > This topic has come up recently due to some potential >> > locking issues (serialized access) when reading the metadata repository >> in >> > a >> > multi-threaded environment. >> > >> >> Do you have any details (JIRA issue?) about the locking issues? We are >> seeing a lot of locked threads on MetaDataRepository.getMetaData under >> load >> with data caching enabled (OpenJPA 1.2.1) > > > No JIRA Issue yet. We're not experiencing deadlocks, we're just > experiencing some serialized access under certain multi-threaded access. > I > haven't pushed on it yet since we just haven't nailed down the exact > scenario yet. Just not enough time... > > >> Is there anything we can do about this? It is reducing our ability to >> scale >> performance. > > > Sorry, I don't know of any work arounds for your situation. And, looking > at > your call stack, it looks to be a slightly different problem. I also > notice > that you are using the WebLogic environment. What version of OpenJPA are > you using? You mention 1.2.1 above, but I thought WebLogic used the 1.1.x > branch of OpenJPA. > > Thanks, > Kevin > > -- View this message in context: http://n2.nabble.com/correct-way-to-load-persistent-metadata-at-startup-tp2865060p3019777.html Sent from the OpenJPA Users mailing list archive at Nabble.com.