mwkirk
commented
11 years ago From sype...@gmail.com on March 14, 2012 16:40:59 Could you please provide a stack trace dump when the CPU is at 100%? Without that information, it is nearly impossible to tell what's wrong. Thank you!
Original author: gonz...@gmail.com (March 12, 2012 15:08:48)
I know this will probably be useless but we used your library with a heavy amount of notifications (more than millions a day).
When we had a more or less a normal amount we were fine, but as we got more and more each day, we noticed an issue with the javapns that locked 1 CPU at 100%, notifications stop working or had a long delay, and it never came back. We reset the server and got it working again. And then the cycle started again. No matter the time of the day, once each day, and with more notifications always sooner, we got 1 CPU at 100%.
It seems like a loop. We replaced javapns with notnoop's library and never got the problem again but most notifications never worked so we are trying to find the problem with javapns.
Here's our code:
PushQueue pool = Push.queue("path", "password", true, 30); pool.start(); ... pool.add(getPayload(notification), notification.getCode_device());
Maybe there's a glitch on the pool implementation. Hope it helps!
What version of the product are you using? On what operating system? Javapns 2.2 latest build JRE 1.6_30 Linux 2.6
Thanks for all of your hard work!
Original issue: http://code.google.com/p/javapns/issues/detail?id=112