I have a problem where hosts are getting wrapped up together in Zabbix. These are two separate boxes, sql1 and sql2, but discovery seems to think they're the same. If you zabbix_get the agent.version, it shows two separate versions. agent.hostname shows the correct hostnames at the correct IPs, and yet we have the two getting wrapped up together. Here's some screenshots of what I mean:
Here's my zabbix_get result. FQDN and IP obscured.
Here's the discovery panel, where sql2 is correlated to sql1, again, identifying info obscured:
And finally, here's the Agent Interfaces section for sql1. Note that sql2 is being viewed as another IP under sql1. Also of note that sql2 *DOES* exist under the hosts tab as its own server. It just also gets monitored as sql1 as well.
Ok, so that's the symptom, here's what I've done to correct:
I've deleted the second agent interface, and made sure both sql1 and sql2 had correct agent information.
I checked /etc/hosts on both boxes to ensure that there's not something funny in there (managed by Ansible, anyways).
I checked a dig on both forward and reverse DNS on watchdog to ensure that we don't have a funny or wrong DNS entry floating about.
I deleted both hosts from the hosts tab and let discovery re-find them.
I'm kind of at a loss, and google doesn't return anything useful that I can find, but maybe I'm using a wonky search string that won't give me what I'm after. Any ideas? As a more minor issue, some hosts in the discovery tab have a space before the IP, like you can see above with sql2 being slightly indented. Any ideas where that's coming from? It's pestering me. >_<
Here's my zabbix_get result. FQDN and IP obscured.
Here's the discovery panel, where sql2 is correlated to sql1, again, identifying info obscured:
And finally, here's the Agent Interfaces section for sql1. Note that sql2 is being viewed as another IP under sql1. Also of note that sql2 *DOES* exist under the hosts tab as its own server. It just also gets monitored as sql1 as well.
Ok, so that's the symptom, here's what I've done to correct:
I've deleted the second agent interface, and made sure both sql1 and sql2 had correct agent information.
I checked /etc/hosts on both boxes to ensure that there's not something funny in there (managed by Ansible, anyways).
I checked a dig on both forward and reverse DNS on watchdog to ensure that we don't have a funny or wrong DNS entry floating about.
I deleted both hosts from the hosts tab and let discovery re-find them.
I'm kind of at a loss, and google doesn't return anything useful that I can find, but maybe I'm using a wonky search string that won't give me what I'm after. Any ideas? As a more minor issue, some hosts in the discovery tab have a space before the IP, like you can see above with sql2 being slightly indented. Any ideas where that's coming from? It's pestering me. >_<