VSAN link to DR shows as dashed line in Fabric Manager

wilson_1234_2 · ‎08-28-2007

We have a 9509 at our HQ site and a 9216 at our DR site.

We have a single VSAN configured from HQ to Dr that accries all of the replication traffic to DR.

We recenlty had a situation where the fcip link to DR was down with:

"2007 Aug 21 06:46:10 lkld-9509-01 %PORT-5-IF_DOWN_TCP_MAX_RETRANSMIT: %$VSAN 50%

$ Interface fcip1 is down(TCP conn. closed - retransmit failure)"

The link has been up fro a week now, but it is showing as a dashed line in Fabric Manager. The connection shows "Up" and all traffic is making it to the VSAN at the DR site.

Before, I could see the link as a solid yellow line when I would look in Fabric Manager.

Anyone have any thoughts on this?

colin.mcnamara · ‎08-28-2007

At some point you lost network connectivity over your FCIP link. You can interpret this message as link failed. The link however will come back up as soon as connectivity is restored (which by your testing it looks like it did).

it is normal for fabric manager to draw a dotted line on fcip or iscsi links. That is how it differentiates between IP and Fibre Channel transport.

If you found this helpful please rate it.

--Colin McNamara

wilson_1234_2 · ‎08-28-2007

I appreciate the reply.

When I lost that link I could see the fcip link flapping.

We have a DS3 to our Dr site and I could not see any reason for the link to fail.

The routeing and connectivity to the remote end was intact.

Have you ever seen the link flap like that and the routing and IP connectivity be ok?

colin.mcnamara · ‎08-28-2007

Not really, most of the time their was an underlying "blip" in the network. It could be spanning tree reconverging in your access layer, routing protocols converging in your provider. Or it could be mice chewing on a cable. Either way.. something interrupted the TCP session between your two devices. I would check your network monitoring logs, and see if you have had any drops.

-Colin

wilson_1234_2 · ‎08-28-2007

Thanks for the reply.

This started at 5:30PM on the 24th and did not stop until 8:30AM the following morning.

Would this issue have been only from MDS switch to switch and in no way could have been related to the SAN?

colin.mcnamara · ‎08-29-2007

It is most likely the switch, its interface, or another device in between.

--Colin

wilson_1234_2 · ‎08-29-2007

Thanks,

Everything seems to be stable now, but is it possible there could be an issue with a port configuration?

I have been here for almost 5 months and it happend one time for those hours I mentioned, nothing since.

I have head that this has happened before.

I have heard that these could be an issue:

Buffer credits - not sure what to do there

Interface not configured for the 45Mbps DS3 link we have.

It seems if those were issues, we would see this more often that once every few months.

Anything I can look for?