PVE 4 KVM live migration problem

[spirit]

Yes it still happening with new deb

Sent from my SM-G900V using Tapatalk

This is strange, it shouldn't happen. (because it don't check the vm config file on resume anymore)

do you tried to restart both hosts ? (or restart pvedaemon and ha-lrm on both nodes)

 

[MasterTH]

got another issue, ha-related but LXC Container

Executing HA migrate for CT 100 to node pve44
unable to open file '/etc/pve/ha/crm_commands.tmp.8611' - No such file or directory
TASK ERROR: command 'ha-manager migrate ct:100 pve44' failed: exit code 2

have a nfs-volume mounted and the CT is running on this


FIXED BY UPDATING TO THE LATEST PACKAGES

 

[alitvak69]

I restarted pve-ha-lrm on all nodes. This problem where ha starts but nothing happens seems to be harder to reproduce but it is not fixed with new patched deb

Sent from my SM-G900V using Tapatalk

 

[alitvak69]

got another issue, ha-related but LXC Container

Executing HA migrate for CT 100 to node pve44
unable to open file '/etc/pve/ha/crm_commands.tmp.8611' - No such file or directory
TASK ERROR: command 'ha-manager migrate ct:100 pve44' failed: exit code 2

have a nfs-volume mounted and the CT is running on this


FIXED BY UPDATING TO THE LATEST PACKAGES

MasterTH,

It seems to be related to problem in a different thread. Unless something was released overnight, your last statement is irrelevant. Please don't hijack the thread.

This is strange, it shouldn't happen. (because it don't check the vm config file on resume anymore)

do you tried to restart both hosts ? (or restart pvedaemon and ha-lrm on both nodes)

Sent from my SM-G900V using Tapatalk

 

[alitvak69]

Spirit,

Forgot to mention. I stopped / started pvedaemon first and then restarted pve-ha-lrm on all nodes.

So no-resume issue doesn't happen anymore. But stuck after HA migrate starts with ha-manager showing vm on source node is still happening. I will run another batch of migrations to see if can reproduce it again for you.

Sent from my SM-G900V using Tapatalk

 

[alitvak69]

I really had to check first . "FIXED BY UPDATING TO THE LATEST PACKAGES" was relevant at least for MasterTH. The new ha related debs were released last night. I just installed them and will report if I can reproduce the HA stuck issue. Interesting enough, these packages don't trigger any pve-ha-lrm or pvedaemon restarts during installation.

 

[spirit]

Interesting enough, these packages don't trigger any pve-ha-lrm or pvedaemon restarts during installation.

I confirm that it's correctly restart, you can check the /var/log/daemon.log

Oct 16 19:41:44 kvmtest1 systemd[1]: pve-ha-lrm.service stopping timed out. Terminating.
Oct 16 19:41:44 kvmtest1 pve-ha-lrm[4267]: received signal TERM
Oct 16 19:42:42 kvmtest1 pmxcfs[1947]: [dcdb] notice: data verification successful
Oct 16 19:43:20 kvmtest1 systemd[1]: pve-ha-lrm.service stop-sigterm timed out. Killing.
Oct 16 19:43:20 kvmtest1 systemd[1]: pve-ha-lrm.service: main process exited, code=killed, status=9/KILL
Oct 16 19:43:20 kvmtest1 systemd[1]: Unit pve-ha-lrm.service entered failed state.
Oct 16 19:43:20 kvmtest1 watchdog-mux[2422]: client did not stop watchdog - disable watchdog updates
Oct 16 19:43:21 kvmtest1 pve-ha-lrm[8568]: starting server

 

[alitvak69]

PVE 4 KVM live migration problem [SOLVED]

Spirit,

Thank you for all your help. You were the only one who took my nagging seriously. So far I couldn't reproduce the problem I complained about.

 

[spirit]

Re: PVE 4 KVM live migration problem [SOLVED]

Spirit,

Thank you for all your help. You were the only one who took my nagging seriously.

No problem. I was glad to help.

[QUOTE ]So far I couldn't reproduce the problem I complained about.[/QUOTE]

Great ! I can't reproduce it anymore too .