5 Apr
2007
5 Apr
'07
12:06 p.m.
On Wed, 2007-04-04 at 16:01 -0400, James Long wrote:
I’ve installed syslog-ng-.2.0.3 on a Red Hat Enterprise Linux 5 server. I’m using eventlog-0.2.5 and libnet 1.1.2.1. I compiled syslog-ng without any problems. I’ve got a fairly generic config file. Whenever I start the service; the process will chew through all of the available ram on the server (2 gb) then terminate with the following error Glib-ERROR gmem.c:514: failed to allocated ### bytes. This usually takes about 15 minutes to happen. Any thoughts?
Are you sure you are not feeding messages back into syslog-ng recursively? How fast does the process grow? How much CPU does the process eat while doing this? -- Bazsi