HP-UX 11.11 and EVA 3000 issue

[Joe Short]

I have a client that recently purchased a new EVA 3000 array, and has SAN attached it to a pair of rp3440 servers. The storage has been configured and presented to the servers. File systems have been created. However, when a server is rebooted, it takes an excessively long time to complete and the attached errors scroll by again and again.

I suspect there is a misconfiguration with the array, but I did not originally configure it, and am not familiar with this array.

[Sandman!]

Looking at your error message it seems like the problem is with the linkage, either between the server and the SAN fabric or between the SAN fabric and the storage array or both. Investigate if the FC cables are good or replace them and try again.

cheers!

[Joe Short]

As near as I can tell, the cables are good.

[Sandman!]

Joe,

Is this an HA environment? and do you have multiple paths to the array LUNs. Also could you post the output of the following commands:

# ioscan -fnC disk
# ioscan -fnC fcp
# ioscan -fnC fc

cheers!

[Joe Short]

Yes, there are 2 HBAs in each server, and 2 switches on the SAN. There are multiple paths and PVLINKS is configured.

[Joe Short]

Is SecurePath still required for multi pathing these arrays with HP-UX?

[Joe Short]

ioscan -fnC disk

[Joe Short]

ioscan -fnC fc

[Joe Short]

ioscan -fnC fcp

[Sandman!]

Joe,

Your "ioscan -fnC disk" shows four paths to each of the LUNs, where in fact you should have only two (two HBAs/server and two SAN switches should give you two paths).

You need SecurePath for an HA environment accessing EVA storage. Installing SecurePath should show only true paths and decrease your reboot time also. After installing SecurePath, increase the PV timeout value from the 30 seconds default to 60 seconds for each of your LUNs.

hope it helps!

[Albert_31]

Hello Joe,

I can think of only the following :-

a) Get the EVA logs checked by hp so that we can rule out any issues with the storage.

b) Check for any zoning related issues on the switch side.

c) This may not be one of the main causes but it is not harmful not to check it..

EVA3k is a active passive controller, I have seen that most of the time the top controller becomes the active one, check from the nsshow of the switch to which port is the primary controller connected and ensure that in the lvm configuration, you have the primary controller luns first. It may help reduce the errors..

[Arunkumar.B]

Hi Joe,


The same problem i experienced in my datacenter.Pls change the time out value for th pv's to 120 using pvchange -t /dev/dsk/cxtyxdz

And then its advised to install securepath for better performance


Cheerrs
Arunkumar

[Joe Short]

The array is on an older firmware. It needs to be updated to version 4.