[Linux-HA] heartbeat cannot read from one of many sockets

[Marcin Przyczyna]

Hello,

on one node of my production cluster running heartbeat 1.0.4
on SuSE SLES8
I get following error messages:

heartbeat: 2005/03/16_18:56:59 ERROR:
ucast: error receiving from socket: Resource temporarily unavailable

This error occurs after I moved the server from one server room
to another. The error occurs on management interface only
(100 Base-TX, no auto-sensing). The exchange of the network card
did not help. The interface shows error rate equal to zero.

As I changed the type of heartbeats from ucast to mcast, 
the connection after init delay has been declared dead.
Ucast heartbeats running on this interface keep it working,
but I get error messages from logsurfer; about
every 10 seconds ;-(

I suppose the node has a problem, which
could be not a real network problem. 

If I work over ssh using management network 
(the suspect one) I get sometimes delays:
the cursor hangs for a moment and than I get every
letter I typed in, like a streak.

Do you have any ideas what a piece of my equipment could
have a malfunction ?

Cheers,
mpr.

-- 
Marcin Przyczyna
Net & Sys Admin,
citiworks AG
mpr at citiworks.de
+49 89 9925 75356