We have test server but how we will create actual load on that. Recreating this problem is a challenge. After restarting, its working fine from last around 3 weeks. One server receives 80,000 call requests in a day on average. Since then no issue.... It is basically serving VXML for telephony calls.
We can upgrade tomcat, Java.... but who will give guarantee this problem wont occur again. I am going clueless. :o And dont tell me please, forget this crash then. I can forget, customer wont. :P Any suggestion? On Thu, Apr 8, 2010 at 3:09 PM, Michael Erskine <mse...@googlemail.com>wrote: > On 8 April 2010 10:27, Shahnaz Ali <shaan20...@gmail.com> wrote: > > Its not possible to do these kind of experiments in running production > > server. But i am sure there is nothing happened like redeploying new war, > > jar or any kind of change. Server was untouched for last several hours. > > Don't you have an identical test/backup server running your production > code? :o > > > The reason of crash even could be out of Log4J, but i m here to discuss, > why > > logs are pointing to only Log4J. And more over all the errors it is > pointing > > should not even work after tomcat restart. Infact everything is working > fine > > after OS restart without any single change. > > Log4j logs it's problems so when underlying problems occur, they get > logged! If your platform (JVM, application server, host OS) is > unstable (or suspect it is), set up some test conditions and try to > recreate the problem. I can't suggest anything more than that without > doing your job for you :) > > > Regards, > Michael Erskine. > > --------------------------------------------------------------------- > To unsubscribe, e-mail: log4j-user-unsubscr...@logging.apache.org > For additional commands, e-mail: log4j-user-h...@logging.apache.org > >
