*ping*
Anyone able to provide any info on this? I am seeing better memory utilization by OJB specifically, which is helpful. I'm still looking for some validation that ObjectCachePerBrokerImpl would be a safe cache to use in a clustered environment. By my reasoning (and my verification that manually executed database changes are immediately visible to OJB), it seems like it should be, but I'm kind of surprised that the clustering HOWTO doesn't mention it.
Thanks in advance,
JoeAt 6:05 AM -0600 3/11/05, Joe Germuska wrote:
Hi:
I'm trying to track down a bit of a memory problem in an application and much of the retained memory tracks back to OJB classes. I'm trying to understand why this is better and what I can do to eliminate this bloat. I also have situations where I have CPU utilization plateau for an extended period.
I'm new to using profilers, but what I think I'm seeing is a high percentage of retained memory tracking back to objects in the cache. I'm wondering if my CPU problems have to do with having a large number of very old items in the cache, causing the GC to have too much work to do when it checks to see what is available to be cleaned up.
From reading http://db.apache.org/ojb/docu/guides/objectcache.html#ObjectCachePerBrokerImpl
I think that this implementation may help me, because no long term caching is done -- the cache is cleared upon every broker close, right? Is this also an adequate solution for running in a clustered environment? If the caches only last for the duration of the persistence activities, then it seems like there are no risks about operations on either side of the cluster operating on divergent cached objects.
I'm going to change the implementation and put this through a QA process (and some more profiling), but I thought I'd see if this resonates with other people or if people have any more insight to share about OJB's caching routines.
To be honest (and if I'm understanding this correctly), it seems like the per broker behavior would be a much more appropriate "default" implementation. I know that it took me rather a while to realize how the default cache was affecting my application, for example, if I had occasion to make manual database changes with SQL.
I look forward to any information people can share about this.
Joe
-- Joe Germuska [EMAIL PROTECTED] http://blog.germuska.com "Narrow minds are weapons made for mass destruction" -The Ex
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
--
Joe Germuska [EMAIL PROTECTED] http://blog.germuska.com "Narrow minds are weapons made for mass destruction" -The Ex
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
