>>> Steven Dake <[email protected]> schrieb am 20.05.2011 um 17:03 in Nachricht
<[email protected]>:
> On 05/20/2011 01:26 AM, Ulrich Windl wrote:
> > Sorry, I had tried the obsolete <[email protected]> before. Again:
> > 
> >>>> "Ulrich Windl" <[email protected]> schrieb am 19.05.2011 
> >>>> um
> > 12:40 in Nachricht <[email protected]>:
> >> Hi!
> >>
> >> In SLES11 SP1 (corosync-1.3.0-5.6.1) one of our two rings is periodically 
> >> set 
> 
> >> FAULTY:
> >>  # grep FAULT /var/log/messages
> >> May 18 13:20:21 host7 corosync[14275]:  [TOTEM ] Marking ringid 0 
> >> interface 
> 
> >> 172.20.16.7 FAULTY - adminisrtative intervention required.
> >> May 18 13:36:32 host7 corosync[14944]:  [TOTEM ] Marking ringid 0 
> >> interface 
> 
> >> 172.20.16.7 FAULTY - adminisrtative intervention required.
> >> May 18 15:18:36 host7 corosync[14944]:  [TOTEM ] Marking ringid 0 
> >> interface 
> 
> >> 172.20.16.7 FAULTY - adminisrtative intervention required.
> >>
> >> I don't know the reason for this, but one effect we see here are packets 
> >> that are occasionally duplicated now and then:
> >> bond0: received packet with own address as source address
> >>
> >> Could this be the reason? If so, can it be made a bit more robust?
> >>
> 
> duplicate packets are ignored by totem

Hi!

OK, so that's not the problem. What about a single lost packet?

> 
> It is hard to tell why your network goes faulty - usually this is caused
> by delays in delivering the token or the token becomes lost.  What is
> your token timeout?

token: 5000
consensus: 6000

> 
> Also, I recommend using 1.3.1 - we have fixed several totem bugs in this
> release.

Because of a support contract I'm restricted to what Novell releases, sorry. If 
you can tell me which issues were fixed, I might suggest an update to Novell.

Regards,
Ulrich


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

Reply via email to