I've opened: https://issues.apache.org/jira/browse/LOG4J2-571
and just closed it. The reason for the CPU thrashing was that it was the AsyncLoggerConfig thread that was going nuts. That needs the property: -DAsyncLoggerConfig.WaitStrategy=Block to be set, not: -DAsyncLogger.WaitStrategy=Block as stated. I'm not sure if both need to be specified - both are available. The last several moths of slowness solved! Yey! -Chris On Wed, Mar 19, 2014 at 9:12 PM, Chris Graham <chrisgw...@gmail.com> wrote: > I've opened: > > > > On Mon, Oct 14, 2013 at 11:52 PM, Chris Graham <chrisgw...@gmail.com>wrote: > >> Oops. >> >> http://pastebin.com/YNny4XTf >> >> :) >> >> -Chris >> >> >> >> On Mon, Oct 14, 2013 at 10:52 PM, Olivier Lamy <ol...@apache.org> wrote: >> >>> On 14 October 2013 11:13, Chris Graham <chrisgw...@gmail.com> wrote: >>> > Hi Olivier. >>> > >>> > I'll try to have a look at that one soon. >>> > >>> > Oddly enough, here is a log4j.xml that is 'broken', in that it is >>> valid XML >>> > (from what I can see) however, there is ZERO output, and the CPU usage >>> is >>> > <1%, which is what it shouldbe. >>> > >>> Mailing list doesn't support attachment. >>> try gist.github.com or http://pastebin.com/. >>> >>> > Is this an issue with the way that log4j parses (and fails for some >>> > reason???) the config file? >>> > >>> > -Chris >>> > >>> > >>> > >>> > On Mon, Oct 14, 2013 at 7:58 AM, Olivier Lamy <ol...@apache.org> >>> wrote: >>> >> >>> >> Hi >>> >> I believe trunk with the last log4j2 beta version should fix that. >>> >> >>> >> Cheers, >>> >> -- >>> >> Olivier >>> >> On Oct 11, 2013 9:33 PM, "Chris Graham" <chrisgw...@gmail.com> wrote: >>> >> >>> >> > Hi All. >>> >> > >>> >> > I've been having no end of fun with the M4 release of Archiva under >>> WAS >>> >> > 8.5 >>> >> > on AIX (64 bit). >>> >> > >>> >> > With archiva doing nothing (ie not processing new releases or >>> serving >>> >> > any >>> >> > requests from any remote maven's), it is consuming anywhere from >>> 40-95% >>> >> > CPU. >>> >> > >>> >> > I managed to get the log4j2.xml file in a state that caused zero >>> logging >>> >> > output, but it also had the added benefit of CPU was around 1%, >>> which is >>> >> > what was expected. >>> >> > >>> >> > So I've managed to get the root logger into debug mode (no other >>> changes >>> >> > to >>> >> > log4j2.xml), and it's sitting on around 40-50% CPU, and all I'm >>> seeing >>> >> > in >>> >> > the log is: >>> >> > >>> >> > 2013-10-11 20:40:21,484 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,497 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,585 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,599 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,686 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,699 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,786 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,800 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,887 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,901 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:21,979 [scheduler1_QuartzSchedulerThread] DEBUG >>> >> > org.quartz.core.QuartzSchedulerThread [] - batch acquisition of 0 >>> >> > triggers >>> >> > 2013-10-11 20:40:21,988 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,001 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,088 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,101 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,189 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,202 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,289 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,302 [Thread-80] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > 2013-10-11 20:40:22,390 [Thread-79] DEBUG >>> >> > org.apache.archiva.redback.components.taskqueue.DefaultTaskQueue [] >>> - >>> >> > pool >>> >> > >>> >> > Is this expected? >>> >> > >>> >> > Any and all comments/help welcome. >>> >> > >>> > >>> > >>> >>> >>> >>> -- >>> Olivier Lamy >>> Ecetera: http://ecetera.com.au >>> http://twitter.com/olamy | http://linkedin.com/in/olamy >>> >> >> >
