-
Incident report
-
Resolution: Duplicate
-
Trivial
-
None
-
7.0.0
-
None
-
Docker, Ubuntu image
We run our Zabbix setup on AWS ECS, with HA disabled. This worked fine on < 7.0 but after upgrading I noticed the following problems:
With HA disabled:
HA manager error: multiple servers have been started without configuring "HANodeName" parameter |
Oddly, only one instance is running at this point. However, it is possible for ECS to have 2 instances of Zabbix-Server active. Once a new instance has spun up, the old one will automatically drain and be removed.
For < 7.0 this worked fine, but with 7.0 this behaviour appears to have changed.
I tried to set
- ZBX_AUTOHANODENAME=fqdn
- ZBX_AUTONODEADDRESS=fqdn
The frontend showed that there were multiple nodes now, one active and one stopped. Looked fine.
However, at some point the process fails again and again restarts the server.
HA manager has been paused Got signal [signal:15(SIGTERM),sender_pid:0,sender_uid:0,reason:0]. Exiting ... |
- duplicates
-
ZBX-24574 HA node flipping between standby and active states
- Closed