NLB Ping question
- From: "Tim Walsh" <tmwalsh@xxxxxxx>
- Date: Mon, 5 Jan 2009 13:19:08 -0500
Has anyone seen this behavior before and is it normal and to be expected?
Could it be associated with the Nodes converging?
We have 2 Windows 2003 Enterprise servers fully patched configured for
Network Load Balancing (NLB) as follows:
Node A:
10.5.105.215
Node B:
10.5.105.216
NLB:
10.5.105.219
When we start a continuous ping on Node A to Node B (10.5.105.216) we'll see
approximately 20% of the pings being dropped. Essentially we'll see the
pings respond within 1ms for 10 minutes, followed by a minute or 2 of no
response. We see the same behaviour from the other Node pinging the first
node at the same time. While the drop outs are occurring we find that we are
able to ping both nodes from client workstations, so the drops appear to be
somehow related to the host servers.
We've checked the logs and find no unusual entries, in either the hardware
or OS logs. We've checked the switches that the servers are connected to and
find no indication of problem on either interface port. We did notice the
MAC address reported for the port doesn't match the one reported on the
server when we run an "ipconfig /all" command, the first 4 numbers are
different, but we suspect this is related to the NLB functions.
Thanks in advance for any insights you might have.
Tim
.
- Follow-Ups:
- Re: NLB Ping question
- From: Russ Kaufmann
- Re: NLB Ping question
- Prev by Date: Re: please help me < creat exchange cluster physical disk>
- Next by Date: Re: move the cluster to different subnet
- Previous by thread: Re: please help me < creat exchange cluster physical disk>
- Next by thread: Re: NLB Ping question
- Index(es):
Relevant Pages
|