Dejan Muhamedagic wrote: > Hi, > > On Mon, Mar 10, 2008 at 09:34:49PM +0100, Johan Hoeke wrote: >> Hi all, >> >> Am testing a 2-node 2.1.3 crm enabled cluster. >> >> Pulled the heartbeat crosslink cable on eth1, to test that the heartbeat >> would continue to be received on eth0 as configured. >> >> from ha.cf: >> bcast eth0 eth1 # Linux >> >> on both nodes: >> [EMAIL PROTECTED] ~]# ps -ef | grep bcast >> root 9377 9374 0 16:44 ? 00:00:00 heartbeat: write: bcast eth0 >> root 9378 9374 0 16:44 ? 00:00:00 heartbeat: read: bcast eth0 >> root 9379 9374 0 16:44 ? 00:00:00 heartbeat: write: bcast eth1 >> root 9380 9374 0 16:44 ? 00:00:00 heartbeat: read: bcast eth1 >> >> Nevertheless, the nodes declare each other dead when the crosslink is >> pulled: >> Mar 10 16:57:21 koch heartbeat: [4648]: WARN: node julia.uvt.nl: is dead >> Mar 10 16:57:21 julia heartbeat: [9374]: WARN: node koch.uvt.nl: is dead >> >> Why is the heartbeat link on eth0 not preventing this? > > No idea. There's nothing in the logs which would indicate where > is the problem. This should work. Did you check with tcpdump if > heartbeats are reaching the other side? Though there are no > error messages about eth0. > > At least this should be easy to debug: run the cluster with only > bcast eth0 and see if that works. You can also try with unicast > and see how that behaves. If the problem is definitely with > heartbeat, you should open a bugzilla. > > Thanks, > > Dejan > >> [hb_report included] >> >> tia, >> Johan
OK, thanks for the quick reply Dejan, i'll try out your suggestions. The heartbeats on eth0 look ok to me judjing from the tcpdump output: [EMAIL PROTECTED] ~]# tcpdump udp port 694 -i eth0 and src koch.uvt.nl tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes 07:35:21.217764 IP koch.uvt.nl.32774 > 137.56.247.255.ha-cluster: UDP, length 215 07:35:22.221588 IP koch.uvt.nl.32774 > 137.56.247.255.ha-cluster: UDP, length 215 07:35:23.225568 IP koch.uvt.nl.32774 > 137.56.247.255.ha-cluster: UDP, length 215 07:35:24.219868 IP koch.uvt.nl.32774 > 137.56.247.255.ha-cluster: UDP, length 217 and [EMAIL PROTECTED] ~]# tcpdump udp port 694 -i eth0 and src julia.uvt.nl tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on eth0, link-type EN10MB (Ethernet), capture size 96 bytes 07:37:46.577258 IP julia.uvt.nl.1030 > 137.56.247.255.ha-cluster: UDP, length 217 07:37:47.581183 IP julia.uvt.nl.1030 > 137.56.247.255.ha-cluster: UDP, length 217 07:37:48.585148 IP julia.uvt.nl.1030 > 137.56.247.255.ha-cluster: UDP, length 217 07:37:49.579107 IP julia.uvt.nl.1030 > 137.56.247.255.ha-cluster: UDP, length 217 I'll try running bcast on eth0 only a.s.a.p. regards, Johan
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Linux-HA mailing list [email protected] http://lists.linux-ha.org/mailman/listinfo/linux-ha See also: http://linux-ha.org/ReportingProblems
