Server denying heartbeat from Proxy

[blakfeld]

So we have a Zabbix server, that seems to work just fine. It used to be linked to a second node, and when it was, everything worked just fine. But we decided that having two nodes was complete overkill, and took down the second one, and just set it up as a proxy, which as far as I can tell is configured properly. However, the server doesn't seem to want to take any communication from it.

On the proxy, I get:

16036:20130105:152723.152 Heartbeat message failed

And on the server:

10488:20130105:153457.363 Heartbeat from active proxy on [129.120.172.68] failed: proxy [dpdcnproxy] not found

Now, my first thought is that the Host isn't properly defined in the zabbix_proxy.conf, however, I copy and pasted from the conf directly into the gui:



I feel like I've been pretty thorough here, and the proxy is clearly trying to talk to the server. Any idea what I've done wrong?

Thanks!

EDIT:
It might be worth mentioning that initially we had two nodes instead of a server and a proxy, so everything on this server is listed as "Node 1", would that cause it to reject the proxy connection?

[blakfeld]

I had tried this previously, but just for the sake of completeness, I'll update this here:

My boss suggested that it might have to be the same as the hostname of the machine, which I had already, tried, but I figured I'd switch it back because that makes a lot of sense to me.

Again, heres the error on the proxy:

 16547:20130105:212115.066 Cannot obtain configuration data from server. Proxy host name might not be matching that on the server.
 16548:20130105:212214.968 Heartbeat message failed

And on the Server:

 16850:20130105:212048.837 Proxy config request from active proxy on [129.120.172.68] failed: proxy [monitor2] not found
 16846:20130105:212148.739 Heartbeat from active proxy on [129.120.172.68] failed: proxy [monitor2] not found

And heres a screen shot of it configured in the Zabbix GUI:

[richlv]

proxy name in frontend (like in your screenshots) must match exactly (case sensitive) "Hostname" value in the proxy config file.

you also must restart the proxy after changing its config file

[blakfeld]

Hmm, I copy and pasted the hostname from the conf file to the frontend, to ensure they were the same. And in the server logs, the hostname appears to be correct when it denies the connection:

 16850:20130105:212048.837 Proxy config request from active proxy on [129.120.172.68] failed: proxy [monitor2] not found
 16846:20130105:212148.739 Heartbeat from active proxy on [129.120.172.68] failed: proxy [monitor2] not found

It's worth noting that it looks like the proxy is trying to talk to the server, and the server is receiving those messages, it's just denying them for whatever reason.

Could it be because the server node was initially set up to be parent to a child node instead of a proxy?

[blakfeld]

Since no one knows, I'm going to rebuild the server, and see if I can get it up and going. I'll come back and post whatever I find!

[blakfeld]

I rebuilt the Zabbix server, and it now communicates with the proxy no problem. I'm not sure exactly where the problem was, but that seems to make it happy.

I suspect it's a problem in the way my Distributed Monitoring was set up, but I'm pulling that out of the air.

Either way, this can be marked as resolved!

[daniel.lin]

I met the same issue, no idea how to fix that without rebuilding the server.

Tried delete and re-add the proxy, rename, etc. No luck so far

[blakfeld]

Unfortunately as you saw, my solution was a little more nuclear.

A little more information, just in case it helps you get on the right path, the cause of my problem seemed to be that Zabbix doesn't like change.

We initially had two servers set up, with "Monitor1" as parent, and "Monitor2" as child. When we nuked "Monitor2" and reconfigured it as a proxy, the server woud not talk to it. It's hard to say if this is because it was still reading as a child node in the database (I removed it from the GUI), or if the fact that Monitor1 was still reading itself as Node1 was the problem.

It seems to me if you ever need to make any real changes to the configuration, you're better off dumping the database and re-installing. However, surely theres a better answer for this kind of thing, Just thought I'd pop in and let you know I never found it

[RHP-997]

I know this thread is old, but I just suffered through a failed migration (old hardware to new hardware in different datacenter) that was caused by proxies not communicating. It seems that in my case, installing bind on the new Zabbix server might have solved the issue. Before installing bind, I tried re-syncing my new copy of all the Zabbix database data, then reinstalling Zabbix on the new server. Neither of which fixed the issue. Then I noticed that one of the logs I was monitoring (I CAN'T REMEMBER WHICH ONE NOW ) was complaining that /var/named/data/named.run didn't exist. So I installed bind. After that... proxy heartbeat succeeds, and server can send configs.

[RHP-997]

I know this thread is old, but I just suffered through a failed migration (old hardware to new hardware in different datacenter) that was caused by proxies not communicating. It seems that in my case, installing bind on the new Zabbix server might have solved the issue. Before installing bind, I tried re-syncing my new copy of all the Zabbix database data, then reinstalling Zabbix on the new server. Neither of which fixed the issue. Then I noticed that one of the logs I was monitoring (I CAN'T REMEMBER WHICH ONE NOW ) was complaining that /var/named/data/named.run didn't exist. So I installed bind. After that... proxy heartbeat succeeds, and server can send configs.

Edit: I installed bind on the Zabbix server, not the proxy.