Re: [Linux-cluster] Quorum Disk and I/O MultiPath problem

2009-04-22 Thread Flavio Junior
On Wed, Apr 22, 2009 at 9:45 PM, Alex Kompel wrote: > It appears that it took 20 sec for path to fail over. quorumd tko is 10 sec > by default. You may want to reduce HBA timeout or tweak tko for quorumd. > Basically you want to set all cluster timeouts to exceed expected failover > time of lower-

Re: [Linux-cluster] Quorum Disk and I/O MultiPath problem

2009-04-22 Thread Alex Kompel
It appears that it took 20 sec for path to fail over. quorumd tko is 10 sec by default. You may want to reduce HBA timeout or tweak tko for quorumd. Basically you want to set all cluster timeouts to exceed expected failover time of lower-level systems. -Alex On Wed, Apr 22, 2009 at 4:31 PM, Flavio

[Linux-cluster] Quorum Disk and I/O MultiPath problem

2009-04-22 Thread Flavio Junior
Hi folks, I'm trying to configure a 2-node cluster using quorum disk as tie-breaker. I'm getting a problem when my active I/O path for quorum disk goes down (I'm testing turn off one (of two) SAN fiber switches), so one node is being fenced. I believe this is not right or can have a better way to