This started after having around more than 500 zabbix process. (I need that many process for polling, unreachable, it's reasonable for our env)
It just kills the main process after 5 seconds and then all the child process get stuck, I'll have to run pkill zabbix_server to get rid of them.
Sometimes it'll die after few more seconds, without the need of pkill process_name. There have been times where timer_process gets zombie'd and zabbix won't start unless I kill it.
It looks it won't be syncing trends and history due to parent process dies before child.
/etc/init.d/zabbix-server stop
[ ok ] Stopping zabbix-server (via systemctl): zabbix-server.service.
7650:20160928:091949.987 Got signal [signal:15(SIGTERM),sender_pid:14000,sender_uid:0,r eason:0]. Exiting ...
14059:20160928:092240.140 Starting Zabbix Server. Zabbix 3.2.0 (revision 62485).
14059:20160928:092240.141 ****** Enabled features ******
When its normal, killing all child, we'll have messages about history and trends sync done.
This is how it should look if it was successful
4010:20160922:040221.941 Got signal [signal:15(SIGTERM),sender_pid:14402,sender_uid:0,r eason:0]. Exiting ...
4010:20160922:040233.559 syncing history data...
4010:20160922:040233.769 syncing history data done
4010:20160922:040233.772 syncing trends data...
4010:20160922:040236.942 syncing trends data done
It just kills the main process after 5 seconds and then all the child process get stuck, I'll have to run pkill zabbix_server to get rid of them.
Sometimes it'll die after few more seconds, without the need of pkill process_name. There have been times where timer_process gets zombie'd and zabbix won't start unless I kill it.
It looks it won't be syncing trends and history due to parent process dies before child.
/etc/init.d/zabbix-server stop
[ ok ] Stopping zabbix-server (via systemctl): zabbix-server.service.
7650:20160928:091949.987 Got signal [signal:15(SIGTERM),sender_pid:14000,sender_uid:0,r eason:0]. Exiting ...
14059:20160928:092240.140 Starting Zabbix Server. Zabbix 3.2.0 (revision 62485).
14059:20160928:092240.141 ****** Enabled features ******
When its normal, killing all child, we'll have messages about history and trends sync done.
This is how it should look if it was successful
4010:20160922:040221.941 Got signal [signal:15(SIGTERM),sender_pid:14402,sender_uid:0,r eason:0]. Exiting ...
4010:20160922:040233.559 syncing history data...
4010:20160922:040233.769 syncing history data done
4010:20160922:040233.772 syncing trends data...
4010:20160922:040236.942 syncing trends data done
Comment