-
Incident report
-
Resolution: Fixed
-
Blocker
-
1.9.5 (alpha)
-
trunk rev 20714
running for a while, one zabbix_server process suddenly is taking quite a lot of memory
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
2859 zabbix 25 5 562m 519m 716 S 0.0 26.0 1:29.96 zabbix_server
unfortunately, server runs with the default logfile size of 1mb, thus startup messages have been long lost. to identify process, strace output is attached. process suspiciously looks like configuration syncer, as it sleeps for 60 seconds in one second increments, then does some selects against the db and semaphore operations, and repeats like that.
also attached is pmap output of this process - there the biggest memory user, unfortunately, is :
0810e000 530836K rw--- [ anon ]
other notable memory users in pmap output can be identified as various server caches. all caching options are at their defaults.
i understand that this most likely is way too little information to figure out what's wrong, so any further hints on how to debug this would be highly appreciated. i'll try running those server processes for a while longer, although system is a bit short on available memory already.