I need help on why we are losing packets between our cluster nodes every
2 minutes on the dot without fail.

We have a clustered service running via an EMC symmetrix, a Fujitsu
Siemens Primepower 650 in 2 different locations and linked together by
DWDM.

The servers are running Solaris 10 and Cluster Service 4.1 for Oracle.
Both the switches and servers are set for 1GB full duplex with auto
negotiation switched off.

Every 2 minutes we are losing packets without fail and we have tried
every solution that we can think off without success.
We have tried multiple ping tests between our servers and without
failure a packet is dropped every 2 minutes.

The packet drops occur on the cluster servers regardless of whether the
cluster services are running or not or if the heartbeats are up or
down.

Initially our customers were continually being thrown out of the oracle
application but this no longer happens due to a modification within the
databases.

If we do the same ping tests on our backup servers which are setup the
same minus the cluster software and heartbeats, no packets are
dropped.

Any help on this matter would be appreciated.


--
camlans
------------------------------------------------------------------------
camlans's Profile: http://unixadmintalk.com/645
View this thread: http://unixadmintalk.com/showthread.php?t=255227