2.7.0 HA-master stops functioning
-
After upgrading from 2.6.0 to 2.7.0 the master node of the cluster would stop functioning at random times. Thereby the backup node would not come up as master. During such event the webinterface and external and internal IP addresses of the master node are not reachable. The problem can be temporarely solved by rebooting the VM. A gracefull shutdown (using Hyper-V) is possible. The problem sometimes happens twice a day, and sometimes not for a week.
Each node is installed as a VM in Hyper-V.
I have also tried to do a fresh install (new VM) of 2.7.0 but the problem remains.
Syslog does not show anything special. Also the metrics in Zabbix do not show anything out of the ordinary.
For now I have reinstalled 2.6.0 -
Do you have any logs from the primary when it does that?
Steve
-
@stephenw10 Sorry, not anymore
-
Hmm, hard to say then. It sounds like the network stack stopped forwarding but the kernel was still running.
I guess I'd try again and grab logs covering the event. If it happens again.
Steve
