- **status**: unassigned --> invalid
---
** [tickets:#2097] Both controllers went for reboot while recovering from split
brain**
**Status:** invalid
**Milestone:** 5.2.FC
**Created:** Thu Oct 06, 2016 04:58 AM UTC by Chani Srivastava
**Last Updated:** Thu Oct 13, 2016 10:11 AM UTC
**Owner:**
To be able to handle split-brain with e.g. stonith there need to be a way for
stonith to communicate with, in this case, the hypervisor, in some other
configuration it may be e..g IPMI. If this only interface has been brought down
stonith can not fence the other node. The correct way is to add
To stimulate split brain scenario we intentionally did not configure redundant
interface.
---
** [tickets:#2097] Both controllers went for reboot while recovering from split
brain**
**Status:** unassigned
**Milestone:** 5.2.FC
**Created:** Thu Oct 06, 2016 04:58 AM UTC by Chani Srivastava
I suggest to close this ticket with status "Invalid". The configuration above
is not correct.
---
** [tickets:#2097] Both controllers went for reboot while recovering from split
brain**
**Status:** unassigned
**Milestone:** 5.2.FC
**Created:** Thu Oct 06, 2016 04:58 AM UTC by Chani
A question, how is the network configured? I assume tipc is used for the
opensaf cluster and there is a separate interface for tcp and stonith, (the
backplane)? How are the interfaces brought down? It seems that even the
"backplane" interface is down.
---
** [tickets:#2097] Both controllers
The command --- virsh --connect=qemu+tcp://192.168.122.1/system list --all
displays
IdName State
26PL-3 running
32SC-1 running
33SC-2
as TCP is used above, check also that /etc/libvirt/libvirtd.conf file, (on the
host):
listen_tcp = 1
is set
---
** [tickets:#2097] Both controllers went for reboot while recovering from split
brain**
**Status:** unassigned
**Milestone:** 5.2.FC
**Created:** Thu Oct 06, 2016 04:58 AM UTC by
I had a quick look at the logs:
Oct 6 10:34:42 SC-1 stonith: [3391]: CRIT: external_reset_req: 'libvirt reset'
for host node failed with rc 1
Oct 6 10:34:42 SC-1 opensaf_reboot: Rebooting remote node SC-2 using stonith
failed, rc: 5
Oct 6 10:34:42 SC-1 osaffmd[1117]: node reboot failure:
---
** [tickets:#2097] Both controllers went for reboot while recovering from split
brain**
**Status:** unassigned
**Milestone:** 5.2.FC
**Created:** Thu Oct 06, 2016 04:58 AM UTC by Chani Srivastava
**Last Updated:** Thu Oct 06, 2016 04:58 AM UTC
**Owner:** nobody
**Attachments:**
-
