On 06/08/2011 09:08 AM, Klaus Darilion wrote:
> Hi!
> 
> I have 2 servers with redundant links: eth0 + eth1
> 
> corosync (1.2.1-4 Debian package) is configured to use 2 interfaces:
> ring0 uses eth0, ring1 uses eth1. rrp_mode: active
> 
> 
> If I unplug eth1, then corosync detects that (corosync-cfgtool -s on
> both servers reports ring1 as faulty).
> 
> However if I unplug eth0, then corosync-cfgtool -s on both servers still
> reports ring0 as "no faults". (However pacemaker detects that the other
> node is offline whereas itself is online.) If I reverse the config so
> that ring0 uses eth1 and ring1 uses eth0, then the same problem happens
> with eth1. Thus, ring0 is always "no faults", even if the Ethernet cable
> is unplugged.
> 
(did you re-nable the ring?)

> Question 1: Is this a bug/feature/know problem that ring0 is always
> reported as "no faults"? Is there a solution/workaround to have proper
> ring status?
> 

see below - file bugs per our support mechanisms

> 
> Next thing is, that if I call "corosync-cfgtool -r" while ring1 is
> faulty, then ring1 status will become "no faults" although it is still
> unplugged - which IMO is another bug. Any reason for this?
> 

see below - file bugs per our support mechanisms
> 
> Would updating to 1.3.0-3 from Debian backports improve the situation?
> 
no

Redundant ring isn't ready for prime time, but it is something that is
in the queue.  see:

https://lists.linux-foundation.org/pipermail/openais/2011-May/016190.html

> Thanks
> Klaus
> _______________________________________________
> Openais mailing list
> [email protected]
> https://lists.linux-foundation.org/mailman/listinfo/openais

_______________________________________________
Openais mailing list
[email protected]
https://lists.linux-foundation.org/mailman/listinfo/openais

Reply via email to