Not quite a hot fix, but fas1/fas2 are currently configured to keep core dumps to aid in debugging recent 500s.
The list of change is:
Created /etc/httpd/conf.d/coredump.conf containing "CoreDumpDirectory /tmp" Edited /etc/profile to not limit core dump sizes: {{{ --- /etc/profile.back 2009-07-10 13:36:53.000000000 +0000 +++ /etc/profile 2009-07-10 13:37:48.000000000 +0000 @@ -27,7 +27,8 @@ fi
# No core files by default -ulimit -S -c 0 > /dev/null 2>&1 +#ulimit -S -c 0 > /dev/null 2>&1 +ulimit -c unlimited > /dev/null 2>&1
if [ -x /usr/bin/id ]; then USER="id -un" }}} Added DAEMON_COREFILE_LIMIT='unlimited' to the end of /etc/sysconfig/init Ran echo 2 > /proc/sys/fs/suid_dumpable (will go away on reboot)
id -un
Most of this apart from the apache config was following the steps at http://kbase.redhat.com/faq/docs/DOC-5353.
Oops, I forgot to mention that this also involved temporarily setting SELinux to permissive on fas1/fas2.
Just did the following on fas1/fas2:
{{{ rm /etc/httpd/conf.d/coredump.conf mv /etc/profile.back /etc/profile remove last line from /etc/sysconfig/init echo 0 > /proc/sys/fs/suid_dumpable /etc/init.d/httpd restart setenforce 1 }}}
so everything is back to normal now. We got some good core dumps and backtraces out of this.
Login to comment on this ticket.