Re: [Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-27 Thread Gianluca Cecchi
unfortunately, doing this seems to have a problematic side effect. Set up -f 1 on one node and -f 10 on the other. Now if I panic one node, it is fenced by the other one, but when restarting it remains in start fencing till it forms after some minute an own cluster and kills cman on the other n

Re: [Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-24 Thread Marc Grimme
On Tuesday 24 February 2009 16:59:26 Gianluca Cecchi wrote: > thanks, but where do I have to put the timeout? > Inside fence seciotn of the nodes: > > > > >

Re: [Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-24 Thread Gianluca Cecchi
thanks, but where do I have to put the timeout? Inside fence seciotn of the nodes: or inside definition of fence devices:

Re: [Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-24 Thread Marc Grimme
We've solved this problem by using fence_timeouts that are dependent on the nodeid. Means node0 gets timeout=0 and node1 gets timeout=10. Then node0 will always survive. That's not the optimum way but works. Or use qdiskd and let it detect the networkpartitioning (whereever it happens) and decid

Re: [Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-24 Thread Gianluca Cecchi
And these are the logs I see on the wo nodes: the first node: Feb 23 16:26:38 oracs1 openais[6020]: [TOTEM] The token was lost in the OPERATIONAL state. Feb 23 16:26:38 oracs1 openais[6020]: [TOTEM] Receive multicast socket recv buffer size (288000 bytes). Feb 23 16:26:38 oracs1 openais[6020]: [TOT

Re: [Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-24 Thread Gianluca Cecchi
Actually my situation is pretty different and worse. two nodes cluster with qdisk and hp ilo based fencing, components rh el 5U3 based. if I panic a node, the other correctly fence it with default action of rebooting it. And also the converse is true. But if for example I get down the intracluster

Re: [Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-24 Thread Kein He
Hi Rajeev, there are several ways to stop the fencing loop: 1. import the third node to the cluster, as a result the quorum votes will great than half total votes . 2. Using qdisk, you can implement the Tie-Breaker IP: use heuristic option to monitor the Gateway. qdisk will also increase the

[Linux-cluster] fencing loop in a 2-node partitioned cluster

2009-02-23 Thread Rajeev P
Hi, I have question regarding a network partition in a 2 node cluster. Consider a 2-node cluster (node1 and node2) setup with a cross-cable for heartbeat and setup to use HP iLO as the fencing mechanism. In the event of network partition (and this case assume that the cross cable was pulled out)