TOPIC:

Hi, guys and girls.
I have 2-node HA cluster of identical HP servers.
Problem started few months ago, do not know exactly when and why.
Almost every time when I login to XenCenter, I get email alert:
"Failed to update global pool configuration settings in /etc/ha-lizard/ha-lizard.pool.conf - Check Configuration!" (check attached screenshot).

We changed our main router last week and its IP address is used for fencing. When I did that, almost all my VM's got their storage corrupted. I had to restore them from backups to be functional again. I do not know if this is related with the alert, but I am a "little" more concerned with HA availability than before.

After that crash, I updated both servers with the latest versions of ha-lizard and iscsi-ha. The error remains.

Thanks for checking this out.

Br, Alko.

The error message is generated when XAPI becomes unresponsive. Given that this is occuring only when connecting with XenCenter, chances are that there is some clock drift between the 2 servers which can cause a significant delay when connecting to XenCenter. Check that the time is in sync between the hosts.

Regarding the router/data corruption issue. Not sure how replacing a router would corrupt your data.

Hi.
Time is identical on both servers.

Regarding router - its IP address is used for fencing/heartbeat. When I disconnected LAN from the first server, the shared storage got somehow lost and all running VM's crashed because they lost their storage also.

The main issue here are those alerts that I am getting at logins.

Hello.
Looks like this problem is getting worse.
I started receiving error messages without me connecting to servers.
Tonight I got those two:
"xe_wrapper: COMMAND: xe pool-list --minimal Has reached the maximum allowable time of 10 seconds. Killing all processes now! "

"xe_wrapper: COMMAND: xe vm-list uuid=9ac2eeff-0aaa-5b5d-9bbd-1f0fe91950e1 Has reached the maximum allowable time of 10 seconds. Killing all processes now! "

I am thinking of reinstalling everything from scratch, including installing 6.5 version of Xenserver.

What do you think? And I am also interested in pro support prices.

Br, Alko.

The alert is being triggered as a result of Xen API hanging. We have seen this in our own testing on older HW with limited memory and CPU.

It is hard to tell exactly why XAPI is hanging in your case. You can try to increase the wait time in the HA-Lizard settings to avoid getting the warnings. Most installations have a default timer of 12 seconds. You can try increasing this to ~20.
From the shell:

"ha-cfg set xe_timeout 20"

Did that, now I have to wait to see if error will come again.
The HW should not be a problem - both servers have 2 x 4 core CPU and 32 GB RAM.

Do you have any instructions for reinstalling HA on 2-server setup where HA was previously installed? Or is it the same as clean install?

Thanks, Alko.