[ovirt-users] hosted_engine, fencing does not work

[martirosov.d]

martirosov.d писал 2015-07-14 17:24:
> Николаев Алексей писал 2015-07-10 10:45:
>> Ovirt больше не поддерживает CentOS 6 для
>> гипервизоров. CentOS 6 поддерживается
>> только для Ovirt Engine.
>> Считаю, что сначало надо
>> воспроизвести проблему на CentOS 7.1. Если
>> она сохранится, то будут нужны логи с
>> гипервизоров (/var/log/vdsm/vdsm.log) и лог Ovirt
>> Engine (/var/log/ovirt-engine/engine.log).
>> 
>> ЛОГИ МОГУТ СОДЕРЖАТЬ
>> КОНФИДЕНЦИАЛЬНУЮ ИНФОРМАЦИЮ О ВАШЕЙ
>> СИСТЕМЕ.
>> 
>> 10.07.2015, 03:06, "martirosov.d" <martirosov.d at emk.ru>:
>> 
>>> Hi.
>>> 
>>> ovirt3.5
>>> 
>>> Have two servers(node1 and node2) that are runnig Centos6.6.
>>> 
>>> 1. engine on node2. Disables network node1, engine restarts after
>>> some
>>> time node1 and everything works fine.
>>> 2. engine on node1. Disables network node1, engine moved to node2,
>>> node1
>>> but does not reset, although attempts.
>>> 
>>> That log messages:
>>> 
>>> 2015-Jul-03, 10:42 Host node1 is not responding. It will stay in
>>> Connecting state for a grace period of 120 seconds and after that an
>>> 
>>> attempt to fence the host will be issued.
>>> 2015-Jul-03, 10:39 Host node1 is not responding. It will stay in
>>> Connecting state for a grace period of 120 seconds and after that an
>>> 
>>> attempt to fence the host will be issued.
>>> 2015-Jul-03, 10:36 Host node1 is not responding. It will stay in
>>> Connecting state for a grace period of 120 seconds and after that an
>>> 
>>> attempt to fence the host will be issued.
>>> 2015-Jul-03, 10:34 User admin at internal logged in.
>>> 2015-Jul-03, 10:33 Host node1 became non responsive. It has no power
>>> 
>>> management configured. Please check the host status, manually reboot
>>> it,
>>> and click "Confirm Host Has Been Rebooted"
>>> 2015-Jul-03, 10:33 Host node2 from cluster emk-cluster was chosen as
>>> a
>>> proxy to execute Status command on Host node1.
>>> 2015-Jul-03, 10:33 Host node1 is non responsive.
>>> 
>>> Manual fencing to node1 works from the node2:
>>> # fence_ipmilan -A password -i 10.64.1.103 -l admin -p *** -o status
>>> Getting status of IPMI:10.64.1.103...Chassis power = On
>>> Done
>>> 
>>> Power Management on node2 test succeeded in oVirt Manager.
>>> 
>>> i.e. If turn off the host on which the engine, after moving to a
>>> healthy
>>> host, engine does not fencing to a problematic host and all VMs that
>>> 
>>> were on problematic host does not migrate.
>>> If the problem is not with the host on which the engine, then it
>>> works:
>>> making fence, and migrates Vms.
>>> 
>>> _______________________________________________
>>> Users mailing list
>>> Users at ovirt.org
>>> http://lists.ovirt.org/mailman/listinfo/users [1]
>> 
>> 
>> Links:
>> ------
>> [1] http://lists.ovirt.org/mailman/listinfo/users
> 
> Здравствуйте.
> 
> Установил CentOS 7.1 на гипервизоры и на engine из ovirt-release35.rpm.
> 
> Ситуация повторилась. Т.е. отключил сеть на node2, на котором был
> engine и тестовая ВМ.
> 
> engine мигрировал на node1, но fence для node2 не сделал и тестовая ВМ
> осталась не доступна и хост node2 тоже.
> 
> Ошибка та же.
> 
> Логи engine и с node1 во вложении.

Если проблемному хосту сделать: Confirm 'Host has been Rebooted', то 
после этого ВМ мигрируют с этого хоста, после этого Maintenance -> 
Activate и хосту делается fence.