I currently have both the server and agent running with logging set to debug level 4. As soon as it dies I will email these over to you Alexei.
Thanks for your help with this.
Thanks for your help with this.
-
-
Any updates on this? This is my favorite thread on the internet ...Comment
-
Server stopped collecting sometime after midnight - by looking at the latest data it apears to be sometime between 00:10 and 00:20 when it died.
Took a little longer to get a log file than expected because the first time round it used up all the space in my /tmp partition! The raw log file is 2GB! I have tarred the file but obviously can't email it because its still 125MB, so I've emailed you a link to download it Alexei. Don't envy you looking through a log so large.
Hopefully we can now work out what is wrong as reading the forums there seems to be quite a few people with a similar problem.Comment
-
Alexei repled:
Thanks for the log files.
Unfortunately I do not see nothing wrong in the log. It seems that ZABBIX was killed, not crashed.
It was stopped exactly at 07:20:00 am. It is very suspicious to me!
Please check your system, it seems that some periodic (?) process killed ZABBIX.
Problem is that the zabbix server is still running (I can see several zabbix processes by doing ps aux at the command line). Any other ideas?Comment
-
Sorry, I just had a thought. 7:20 is when I restarted the service myself to get it going again. It stopped collecting data around midnight.Comment
-
Comment
-
I'm assuming this means you found something in the logs? If so, I for one and infinity005 will be very happy
Comment
-
Yes, I found something! Actually this was an known problem, which suddenly came up. The problem affects all ZABBIX systems, especially those with heavy use of active checks for monitoring of unreliable networks and remote locations.Comment
-
Comment
-
The patch is attached for your convenience. It has been tested!Comment
-
Check out ZBX-343 in svn. I think its the patch.
Installed it on my 1.4.4 setup. FINGERS CROSSED!!Last edited by bbrendon; 25-03-2008, 19:08.Comment
-
I was looking at ZBX-323.
Can you tell me how to install a patch?Comment
-
I've managed to install the patch.
I'll let you know how it goes - but I too will sit with my fingers crossed. If it makes it through the night I'll be happy, probably won't be convinced until its run for a month though
Comment
-
No need to wait a month. Feel free to keep me updated every new week of ZABBIX uptime
Comment
-
My zabbix server just crashed.
Its 1.4.4 + load patch and trapper patch. It was fine with the load patch, except that it would suddenly hang. The trapper patch causes it to crash.
I'm upgrading to 1.4.5 and bumping debugging back up to 4. I have the 1.4.4 + patch, debug=3 log if you're interested with summary below:
Code:# tail zabbix_server.log.crash1.144_patched ] is not suitable for [[email protected]] 32331:20080325:150003 Expression [{21451}>50] cannot be evaluated [Unable to get value for functionid [21451]] 32331:20080325:150003 Expression [{21456}>50] cannot be evaluated [Unable to get value for functionid [21456]] 32331:20080325:150003 Expression [{21457}>50] cannot be evaluated [Unable to get value for functionid [21457]] 32322:20080325:150020 Active parameter [system.run[mysqladmin --defaults-file=/etc/zabbix/agent.mycnf status|cut -f4 -d":"|cut -f1 -d"S"]] is not supported by agent on host [arts.web2] 32319:20080325:150021 One child process died. Exiting ... 32319:20080325:150023 ZABBIX Server stopped # grep 32319 zabbix_server.log.crash1.144_patched 32319:20080325:150021 One child process died. Exiting ... 32319:20080325:150023 ZABBIX Server stopped #
Last edited by bbrendon; 26-03-2008, 01:14.Comment
Comment