Host status unknown... forever

[osoh]

Ok, I've read some posts about similar issues in the forum, but I can't find a solution.

I've got some servers (that inherits from a common template) which status remains "unknown", even when all the items seems to work pretty fine.

The problem started when I tweaked the trigger "{HOSTNAME} is unreachable" to add some "timeout". But I've come back to the old expression and it doesn't work now (always show "unknown").

I've deleted one of those hosts, created it again, linked with the same template and it works now! (only that host). But re-creating the hosts is not a solution for me, because we've got more than 800 devices being monitorized.

Any idea?

[knarfling]

One of the things on my system (1.4.2) is that the status item only updates when there is a change in status. In other words, if it starts in unknown when the server is up, it stays in unknown forever.

The only way I found to update it was to change the status of the server. For production servers I moved the host to a maintenance group (my actions are based on group so I don't get tons of pages for a planned outage) and then shut down the zabbix agent on the specific machine. I waited until it registered on zabbix server, then restarted the agent.

When installing a new server, I now start the agent, get everything set, stop the agent and restart it before putting it into production. That way the status is set properly all the time.

There may be other ways to change it, but this one works for me.