I didn't see any others posts on this, so maybe we're a unique flower, but we were seeing issues with the default "Template App MySQL", where a stopped mysql server would simply return no data for the MySQL status (mysql.ping) item and therefore the "MySQL is down" trigger would not alert properly.
Best I can tell this was related to the fact that there was a bunch of stderr spit out by mysqladmin when the actual service was stopped (i.e. not "failed") and this caused the item to not update since it was expecting a decimal value (0 or 1) and got a bunch of text.
We solved it by updating the "mysql.ping" item (in userparameter_mysql.conf) from:
UserParameter=mysql.ping,HOME=/var/lib/zabbix mysqladmin ping | grep -c alive
to:
UserParameter=mysql.ping,HOME=/var/lib/zabbix mysqladmin ping 2>&1 | grep -c alive
(i.e. redirecting stderr to stdout _before_ the grep)
We're running zabbix 2.4.4 server/agent
Server on ubuntu 14.04 and agents are 12.04(mostly) / 14.04
Hope it helps someone.
Best I can tell this was related to the fact that there was a bunch of stderr spit out by mysqladmin when the actual service was stopped (i.e. not "failed") and this caused the item to not update since it was expecting a decimal value (0 or 1) and got a bunch of text.
We solved it by updating the "mysql.ping" item (in userparameter_mysql.conf) from:
UserParameter=mysql.ping,HOME=/var/lib/zabbix mysqladmin ping | grep -c alive
to:
UserParameter=mysql.ping,HOME=/var/lib/zabbix mysqladmin ping 2>&1 | grep -c alive
(i.e. redirecting stderr to stdout _before_ the grep)
We're running zabbix 2.4.4 server/agent
Server on ubuntu 14.04 and agents are 12.04(mostly) / 14.04
Hope it helps someone.