Re: two clusters - same subnet - missing packets

At 02:47 PM 6/1/2005 +0100, you wrote:
>Hi All,
>
>I have two clusters one for postfix - other for dns on same subnet,
>
>I was getting on all 
>
>heartbeat: 2005/06/01_14:03:44 info: MSG: Dumping message with 9
>fields
>heartbeat: 2005/06/01_14:03:44 info: MSG[0] : [t=status]
>heartbeat: 2005/06/01_14:03:44 info: MSG[1] : [st=active]
>heartbeat: 2005/06/01_14:03:44 info: MSG[2] : [src=obedns2]
>heartbeat: 2005/06/01_14:03:44 info: MSG[3] : [seq=8e]
>heartbeat: 2005/06/01_14:03:44 info: MSG[4] : [hg=1]
>heartbeat: 2005/06/01_14:03:44 info: MSG[5] : [ts=429db175]
>heartbeat: 2005/06/01_14:03:44 info: MSG[6] : [ld=n/a]
>heartbeat: 2005/06/01_14:03:44 info: MSG[7] : [ttl=3]
>heartbeat: 2005/06/01_14:03:44 info: MSG[8] : [auth=1 b0da8776]
>heartbeat: 2005/06/01_14:03:44 ERROR: process_status_message: bad node
>[nodename blah] in message
>

the [src=obedns2] is the source node; It complains a mismatch from the source node to the node name you put in ha.cf

make sure your node name in ha.cf is the same as 'uname -n'

>assumed was due to multicasts in all my configs in my ha.cf so have
>changed all to ucast
>
>so
>
>now have
>
>ucast eri0 (ip of second )
>
>and
>
>ucast eri0 ( ip of first )
>
>and all seems fine,
>
>however im getting the odd missing packet
>
>eartbeat: 2005/06/01_14:24:57 WARN: 1 lost packet(s) for [obedns2]
>[86:88]
>heartbeat: 2005/06/01_14:24:57 info: No pkts missing from obedns2!
>heartbeat: 2005/06/01_14:33:20 WARN: 1 lost packet(s) for [obedns2]
>[336:338]
>heartbeat: 2005/06/01_14:33:20 info: No pkts missing from obedns2!
>
>
>same subnet though - so rule out routing issues
>
>version 1.2.2 on all nodes

If this happens only couple of times, it's ok. Lost messages are recovered by heartbeat.

-Guochun

_______________________________________________
Linux-HA mailing list
Linux-HA <at> lists.linux-ha.org
http://lists.linux-ha.org/mailman/listinfo/linux-ha