Thanks for your inputs. I won't go into all the business details, but
production and degraded service is two different things. Anyway :)
I think I resolved my issue. The LACP timeouts were set to long on the
hypervisor and the switches. I changed that to short, and no more flip
on a switch failure.
Thanks!
FG
On 1/23/2014, 3:29 AM, Lars Marowsky-Bree wrote:
On 2014-01-22T15:02:11, Francois Gaudreault <fgaudreault@xxxxxxxxxxxx> wrote:
Well LACP is at the hypervisor level, so for Corosync, it's a standard
interface.
Active/Passive is not really an option for us, we need the 2GB bandwidth.
Any timeouts you think we can tweak?
Hi all,
while I agree with digimer's assessment, the last paragraph above is a
bit tricky. Because if you *need* the 2GBit/s bandwidth (in order to
satisfy the business demand), you cannot really sustain your services
during degraded mode - which is the whole point of bonding, right? ;-)
Best,
Lars
--
Francois Gaudreault
Architecte de Solution Cloud | Cloud Solutions Architect
fgaudreault@xxxxxxxxxxxx
514-629-6775
- - -
CloudOps
420 rue Guy
Montréal QC H3J 1S6
www.cloudops.com
@CloudOps_
_______________________________________________
discuss mailing list
discuss@xxxxxxxxxxxx
http://lists.corosync.org/mailman/listinfo/discuss
