HPE Storage Users Group

I have a 3Pare F200 3.1.2 (MU5) and a 7200 3.1.2 (MU1) that recently started having nodes panic when under high RCIP load. They are both targets for multiple RCIP groups that have been fine for a year+. We are now seeing more frequent RC link down errors and occasionally a node crashes and reboots. This started after network switch IOS updates were performed but there are no other issues that have come about since the update for numerous servers, apps and SANS running over the same links that were updated. HP has struggled to get back anything other than update to MU5 which was done on the F200 but there have been subsequent node panics since the upgrade. They have also mentioned setting the RCIP and Admin ports to Auto but with no real strong reason why.

Today I saw an article on this site about a similar issue with a 3.1.3 unit having panic's caused by an OS bug as a result of being set to 1000 Full and not Auto but why would this happen suddenly? The F200 is directly plugged into the switches that had the IOS upgraded, so I can see that as being a possible cause, but the 7200 is plugged into a different set of switches that have not been touched in over a year. The switches the 7200 ports are plugged in, however are uplinked to the switches that were upgraded.

Any thoughts or input would be greatly appreciated.

Are your rcip ports in fact set to auto/auto?

I have a pair of 7200's on 3.1.2 MU2 and have no issue even under high load.

In fact tonight I just dumped 600gb onto a remote copy vv and had no issue

Edited:damn you autocorrect!

The ports are all currently set at 1000 full not auto. We have been replicating about 15 TB and growing to the F200 asynchronously for over a year and probably 12 TB to the 7200.

I would pause the replication and set them to auto...

HP has had a few issues with ports not being auto, so between that and the fact that it's really a minor change that should be painless and quick, I would do it just for shits and giggles and see what happens.

Can't make it any worse, right?

That's the plan tomorrow. Just wish HP could offer some docs to back their Concern.

Thanks
MJK

I've had that same issue with them.

In some ways I think cisco does it better. Every bug has a specific csc number and document.

Hp would do well to do likewise