[pve-devel] Proxmox 4 feedback

[Gilou]

Le 09/10/2015 19:01, Gilou a écrit :
> Le 09/10/2015 18:43, Thomas Lamprecht a écrit :
>> Whats your cluster setup? Three nodes?
> 
> 3 nodes, yes
> 
>>
>> Output from:
>> # ha-manager status
> 
> root at pve2:~# date
> Fri Oct  9 19:00:32 CEST 2015
> root at pve2:~# ha-manager status
> quorum OK
> master pve2 (old timestamp - dead?, Fri Oct  9 18:57:03 2015)
> lrm pve1 (active, Fri Oct  9 19:00:30 2015)
> lrm pve2 (active, Fri Oct  9 19:00:31 2015)
> lrm pve3 (active, Fri Oct  9 19:00:30 2015)
> 
> I removed the service manually, I'm going to start again and record it...
> 
>>> They were down (so I guess the fencing worked). But still?
>> They were down means they rebooted during the time?
>> If the services are frozen it means for sure that a gracefull shutdown
>> of the LRM happened, that normally also comes together with a gracefull
>> shutdown of the whole node.
> 
> Nothing happened but the cables being unplugged.
> 

OK, well, gonna start fresh, dunno what is happening:
failed state.
Oct 09 19:05:49 pve2 watchdog-mux[1619]: watchdog set timeout: Invalid
argument
Oct 09 19:05:49 pve2 systemd[1]: watchdog-mux.service: main process
exited, code=exited, status=1/FAILURE
Oct 09 19:05:49 pve2 systemd[1]: Unit watchdog-mux.service entered
failed state.
Oct 09 19:05:49 pve2 watchdog-mux[1621]: watchdog set timeout: Invalid
argument


But it's the second time I try this, and it is inconsistent, and not
really pleasant. I'll get back with full install of PVE 4 as of the
release, and try that again. Stay tuned...