1.8.10, discovery, web & manually entered hosts not working??

[magic-man]

Hi Everyone,

I recently did a clean install of 1.8.10 on my development servers instead of upgrading 1.8.6.

Zabbix Server, PostGre & Apache all have their own servers.

I've verified zabbix server can connect to the database by checking for active connections inside postgre, and the webUI can talk to the database & postgre server.

However the check I added for my website and discovery aren't working. I decided to try helping it out and added these 3 servers manually, but I'm still not getting any data.

The only things I see in the zabbix server log are:

1222:20120106:152352.974 [Z3001] connection to database 'zabbix' failed: [0] could not connect to server: Network is unreachable
	Is the server running on host "zxs.incite-edge.org" and accepting
	TCP/IP connections on port 5432?

  1222:20120106:152352.974 Cannot connect to the database. Exiting...
  1216:20120106:152418.401 Starting Zabbix Server. Zabbix 1.8.10 (revision 24303).
  1216:20120106:152418.404 ****** Enabled features ******
  1216:20120106:152418.404 SNMP monitoring:           YES
  1216:20120106:152418.404 IPMI monitoring:           YES
  1216:20120106:152418.404 WEB monitoring:            YES
  1216:20120106:152418.404 Jabber notifications:      YES
  1216:20120106:152418.404 Ez Texting notifications:  YES
  1216:20120106:152418.404 ODBC:                       NO
  1216:20120106:152418.404 SSH2 support:              YES
  1216:20120106:152418.404 IPv6 support:               NO
  1216:20120106:152418.405 ******************************
  1216:20120106:152418.405 NodeID:                      1
  1216:20120106:152418.405 ******************************

I'm not sure why it says it can't access the database when I know it is, I can see 47 connections to the database when the webUI is off. Maybe zabbix goes up before networking is done??

Anyway, it then goes through listing the daemons starting then a bunch of these:

  1269:20120106:155105.577 Sending list of active checks to [10.10.10.71] failed: host [zxs.incite-edge.org] not found
  1269:20120106:155118.784 Sending list of active checks to [10.10.10.72] failed: host [zxp.incite-edge.org] not found

It had these BEFORE I manually added the servers so I think the agents on both are talking to the zabbix server, it just can't send back.

IPtables is down (for now), SELinux is off, their hostnames resolve in my DNS server correctly, and neither IP or hostname as "name" in the GUI seems to work.

Any suggestions would be very greatly appreciated.

EDIT:

This is how it is currently, where the database server is trying to use it's IP, and the zabbix server is trying to talk to itself using it's hostname (FYI, 127.0.0.1 didn't work either).

  1237:20120106:161532.926 Sending list of active checks to [10.10.10.72] failed: host [zxp.incite-edge.org] not found
  1237:20120106:161644.633 Sending list of active checks to [10.10.10.71] failed: host [10.10.10.71] not found

[walterheck]

* The host not found error means there is no host defined in the web frontend with that name
* The fact that you see postgres conections when logged into the frontend doesn;t mean teh zabbix binary can reach your database. In fact, log out of the frontend and check the connection between postgres server and zabbix server first

good luck!

[magic-man]

That's why I'm confused.

I hate repeat questions so I scoured the web for a week before posting. I've read through the forum and found the name in GUI and agentd.conf all need to be the same so I've checked 8 times they are in sync.

I even changed the database server to use it's IP instead and switched the agentd.conf and hostname in GUI, no change.

As the screenshots show, the name in UI matches the names in the error saying it can't find the host.

Edit:

Sorry, to address the database connection between zabbix server and postgres. I mean I can start the database server and see 0 connections. Boot the Zabbix server and see 47 connections to the database when logged into postgres, and another connection when web server is booted.

[magic-man]

Now even more confused

So I just finished reinstalling the Zabbix server to see f it was something stupid I did, but left the webUI and database servers alone.

After reinstalling the server and agent on that machine. I cleared manually entered hosts. The zabbix_server.log continues to say it can't reach 10.10.10.71 and 10.10.10.72.

So out of curiosity, I changed the names in the agentd.conf files. The error's instantly switched to the new host names.

What does that mean? Clearly the agents can talk to the server, and the webUI and server can talk to the database since I can see active sessions with only the webserver or zabbix server running.

Is it a permissions issue somewhere?

[magic-man]

Turns out the issue was that some of the database tables ended up being owned by postgres instead of zabbix which it didn't like.