3.6 KiB

Raw Blame History

Troubleshooting

Finding the active Zabbix server node

To find the node(s) where Zabbix server is active, run the following command on Fuel master node:

# fuel nodes | grep controller | awk -F\| '{print $1,$NF}' | sort -n -k 2 | \
    uniq -s 1 | while read cnode lenv; do echo "=========== Environment $lenv" ; \
        ssh -q node-$cnode 'for r in p_zabbix-server vip__public; do \
        crm resource status $r; \
        done' ; done
=========== Environment 1
resource p_zabbix-server is running on: node-4.test.domain.local
resource vip__public is running on: node-3.test.domain.local

Finding the public VIP

On the returned node from the command above for a given environment, you might also want to know what is the Zabbix VIP address, so run the following command on Fuel master node:

# ssh -q node-3 ip netns exec haproxy ifconfig b_public | \
    grep 'inet addr:' | sed -e 's/[^:]\*://' -e 's/ .\*//'
172.16.0.2

Finding the management VIP

On the returned node from the command above for a given environment, you might also want to know what is the Zabbix VIP address, so run the following command on Fuel master node:

# ssh -q node-4 ip netns exec zabbix ifconfig b_zbx_vip_mgmt | \
    grep 'inet addr:' | sed -e 's/[^:]*://' -e 's/ .*//'
192.168.0.3
# ssh -q node-4 awk '/zbx_vip_mgmt/ {n=1} n==1 && /ipaddr/ {print;exit}' \
    /etc/astute.yaml | sed -e 's/.*: //'
192.168.0.3

Connect to Zabbix Web GUI

Use the URI using the public VIP:

http://172.16.0.2/zabbix

If you cannot access to the Zabbix UI, check that the HTTP Apache server is running on all controller nodes:

# /etc/init.d/apache2 status
* apache2 is running

Zabbix server

If the Zabbix UI reports 'Zabbix server is not running', check the following:

Check if the zabbix-server process runs and where is located, in the following example the server runs on node-2:
```
# crm status
  [snip]
p_zabbix-server      (ocf::fuel:zabbix-server):    Started node-2.test.domain.local
```
Check logs in '/var/log/zabbix/zabbix_server.log' to see eventual error.
If the zabbix-server is down, start it by using the pacemaker command:
```
# crm resource start p_zabbix-server
```

If the zabbix-server is still down, try the following:

# crm resource stop p_zabbix-server
# crm resource cleanup p_zabbix-server
# crm resource start p_zabbix-server

If after the previous commands the zabbix-server is still down and you didn't find any explanation in the logs, try to increase the log level:
```
# sed -i 's/DebugLevel=3/DebugLevel=4/' /etc/zabbix/zabbix_server.conf
# crm resource restart p_zabbix-server
```

Zabbix agents

If a Zabbix agent don't report data (this can be determined on the Zabbix UI page: configuration > hosts).

Check if the corresponding agent is running:
```
# /etc/init.d/zabbix-agent status
```
Restart the zabbix-agent if not running:
```
# /etc/init.d/zabbix-agent restart
```
If the zabbix-agent is still down or doesn't report any data try the following command to validate the agent's configuration. This command should display all data that agent is configured to collect, if not the command should display an explicit error with regard to the configuration:
```
# zabbix_agentd -p
```

Zabbix log files

On any of the cluster node, you might want to look into the Zabbix agents and server log files under:

/var/log/zabbix

3.6 KiB Raw Blame History