Quote:
Originally Posted by mmorse  |
It seems that there is a fix in progress for this bug. Meanwhile, I believe that I have hit it in my personal Open Source installation. Never thought I would, as it only has a few domains with a few accounts (really little load), and it's plenty of anti-spam filters that prevent most of spam from ever reaching the point in which SpamAssassin and ClamAV have to do their bit of work.
Anyway, the VPS (Xen) where the server is running in had out of memory errors. I am not entirely sure whether that was purely due to that bug or something else (the VPS has 1 GB of RAM). Here you are some bits of logs which I find relevant. Please let me know if posting some more can be of any help stating the Java bug caused the out of memory error or it was something else (like the stats generation processes).
See attached file for /var/log/syslog output. You'll see the gap from Jul 22 22:00:04 until Jul 23 00:12:38 (the moment at which the machine was restarted).
According to the log, do you believe it was the Zimbra Logger? Also, I am seeing those ext3 fsck inode cleanup lines at the end of the kernel restart and they are making me quite nervous, too
Thanks.