Ad Widget

Collapse

Zabbix server is not running | Zabbix 4.2 HA with pacemaker

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • zabbix_noobie
    Junior Member
    • Sep 2019
    • 3

    #1

    Zabbix server is not running | Zabbix 4.2 HA with pacemaker

    Hi All,

    Pretty new to this zabbix, my company needs HA for the zabbix servers. So I followed this guide and managed to set up the pacemaker.

    I followed this guide http://yallalabs.com/linux/how-to-co...1#comment-7359

    I configured step by step and used this command to create VIP

    pcs resource create cluster_vip ocf:heartbeat:IPaddr2 ip=127.0.0.2 cidr_netmask=24 op monitor interval=20s

    I can access zabbix dashboard on 127.0.0.2, however I am facing two problems.

    1) Since zabbix agent is running on both master 1 and master 2, its creating conflict in DB. I have pointed both zabbix servers to the same DB.

    PGRES_FATAL_ERROR:ERROR: duplicate key value violates unique constraint “events_pkey”


    and I see this error on dashboard ” Zabbix server is not running: the information displayed may not be current”


    2) What IP entry do I make in my zabbix agents? They are not be able to pick up zabbix servers since its running on 127.0.0.2 and it fails to even telnet on the master server.

    I hope if someone could help me out with this HA.


    My question is, lets say my main zabbix server is zabbix_master1 and secondary is zabbix_master2, I want to use zabbix_master2 as standby, how do I make it work?
  • zabbix_noobie
    Junior Member
    • Sep 2019
    • 3

    #2
    I managed to make the failover work. however two problems still exists

    1) What IP should I use on the agent? If I use the cluster IP it does not work

    2) Web servers are still showing that zabbix server is not running.

    Comment

    • ITOMDave
      Member
      • Nov 2018
      • 53

      #3
      Hi,
      The yallalabs article is a great guide, but treat it as a guide rather than as a detailed recipe.

      The first comment I'd make is that you can only have 1 active Zabbix server at any point in time. i.e. An active-passive cluster rather than an active-active cluster. Assuming this is how you have confgured it that way then that's a great start point.

      The second comment is that you absolutely should have the CLUSTER IP as the address in the agent. If it's not working on the cluster ip I'd first check whether SELINUX is interfering. You can do that by using the "sestatus" command. If sestatus returns "enforcing" you might want to try running "setenforce permissive" and retesting.

      Let us know how you get on.

      Dave@Nobius

      Comment

      Working...