Dejan Muhamedagic wrote:
On Thu, Jun 21, 2007 at 12:38:24PM -0400, Madison Kelly wrote:
Dejan Muhamedagic wrote:
This is probably a long shot, but could you try removing the
serial option from your configuration and see if it makes messages
go away. Otherwise, this looks so serious that I doubt that it
could be the making of the heartbeat.
Thanks.
On a fresh install of Debian Etch using the SUSE-built debian package:
http://ftp.ale.org/pub/mirrors/opensuse/opensuse/repositories/server:/ha-clustering/Debian_Etch/i386/heartbeat_2.1.0-1_i386.deb
and with the serial line commented out the errors still occur.
If this isn't the making of heartbeat, what could it be? As soon as I
stop the heartbeat service, the errors stop being generated. Moreso, the
errors have the 'heartbeat' name in them.
Sorry, I was not clear enough. It is coming from heartbeat. What I
meant is that there's probably something else on your systems
which is causing this to happen. That messaging code is, I
believe, old and should've been well tested by now. Otherwise,
there would have been tons of reports. Or you were really lucky
and stepped on a new bug ;-)
Could this be a problem with something in my system that isn't used
except by heartbeat? Would it help if I recompiled the kernel with
certain options enabled or disabled?
I doubt that it is the kernel.
At this point I am almost ready to try anything...
I suppose that Alan should be able to advise on the matter.
I just got a note from Lars via Bugzilla that it was probably the remote
servers I was pinging mangling the ICMP packets. I changed the 'ping' in
ha.cf to the local firewall and the errors stopped printing.
So it's not so much a bug as unclear handling of a problem? Perhaps this
is because my servers are behind a NAT'ing iptables firewall? Either
way, I am grateful for everyones help and can now *finally* move on with
DRBD!
Madi
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
