jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Steven Githens <swgit...@mtu.edu>
Subject Re: CacheManager slowing performance
Date Fri, 30 Nov 2007 03:37:42 GMT
I'm actually having size-able problems with this right now, any pointers 
would be really appreciated.  (sorry for the pun :)   ).

I'm working on our migration tools to move data from an old legacy store 
to JCR.  Which basically involves just moving over thousands of files 
and directories.  I basically get output like below, where fairly 
frequently, it just hangs on the resizeAll, sometimes for a full minute 
or two.  ( Occasionally, my db pool gets exhausted too, I'm pretty sure 
I'm closing all my sessions ).  This happens fairly frequently ( I've 
been testing with like a  thousand or so nt:files/nt:folders).

I tried turning off my cluster settings since I'm on a single node at 
the moment.  Before that it would just stop on some journal logging for 
a minute or two at a time occasionally.

Is there a penalty for rapidly opening and closing sessions?  Should I 
try to copy N number of files before getting a new Session?

Any suggestions on basically rapidly inserting tons of data for 
something like a migration would be appreciated.

Thanks,
Steve


INFO: resizeAll size=8 (2007-11-29 18:36:48,832 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:36:58,931 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:00,014 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: merged 433 documents in 342 ms into _g. (2007-11-29 18:37:04,731 
IndexMerger_org.apache.jackrabbit.core.query.lucene.IndexMerger)
INFO: resizeAll size=8 (2007-11-29 18:37:10,412 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:11,590 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:22,184 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:23,218 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:24,405 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:35,518 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:39,613 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:40,635 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:50,727 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)
INFO: resizeAll size=8 (2007-11-29 18:37:51,775 
Timer-16_org.apache.jackrabbit.core.state.CacheManager)


malsi wrote:
> I have a which is that from time to time  the class
> org.apache.jackrabbit.core.state.CacheManager gets fired and this slows
> performance very much. . I know that CacheManager  manages the size of 
> the
> caches used in Jackrabbit (which has to be dynamic to increase 
> performance
> and memory usage) but is there another way to avoid having this 
> CacheManager
> class getting fired all the time i.e (schedule it).
>   




malsi wrote:
> I have a which is that from time to time  the class
> org.apache.jackrabbit.core.state.CacheManager gets fired and this slows
> performance very much. . I know that CacheManager  manages the size of the
> caches used in Jackrabbit (which has to be dynamic to increase performance
> and memory usage) but is there another way to avoid having this CacheManager
> class getting fired all the time i.e (schedule it).
>   

Mime
View raw message