Hello,
we run several agents in combination with multiple instance mysql on a single server. (One agent per instance). For the last couple of days we repeatedly face the problem that a few of the agents just die and they complain that the server monitored has "gone away". But this is definitely not the case, the monitored mysql servers were not even restarted or anything for weeks. It also tells to grant some permission which were already granted.
This is what comes up in the log:
2011-02-28 19:23:32: (critical) agent_mysqld.c:723: successfully connected to database at 192.168.40.191:3322 as user mysqlma (with password: YES)
2011-02-28 19:32:33: (critical) agent_mysqld.c:723: successfully connected to database at 192.168.40.191:3322 as user mysqlma (with password: YES)
2011-02-28 20:05:30: (critical) agent_mysqld.c:723: successfully connected to database at 192.168.40.191:3322 as user mysqlma (with password: YES)
2011-02-28 20:06:05: (critical) last message repeated 1 times
2011-02-28 20:06:05: (critical) network-io.c:269: curl_easy_perform('
https://epagesagt@dil.store:18443/heartbeat';) failed: Connection time-out (curl-error = 'Timeout was reached' (28))
2011-02-28 20:06:06: (critical) network-io.c:311: successfully reconnected to dashboard at
https://epagesagt@dil.store:18443/heartbeat
2011-02-28 20:06:35: (critical) agent_mysqld.c:512: [192.168.40.191:3322] mysql_real_query("SELECT value FROM mysql.inventory WHERE name = 'uuid'") failed: Lost connection to MySQL server during query (errno=2013)
Please GRANT SELECT ON mysql.inventory TO 'mysqlma'@'%';
This necessary for correct execution of the agent. Shutting down now.
After restarting the agent everything is fine again. I'm thinking about fixing the start skript to automatically restart, but I don't know wheter this is wise or not. Any hints to what this problem might be related to? Would it help to upgrade to the newest version? (current version is 2.2.3.1755-linux-glibc2.3-x86-64bit, the newer 2.3.1 is available).