IBM LUNs removed by MPIO when attaching NetApp LUN

[steven_jenner]

Hi, yesterday I created and presented a NetApp LUN to a Windows 2008 server that was already attached to some IBM LUNs presented via IBM SVC with a DS3400 SAN at the back. The purpose of this was to enable the migration of data from the IBM storage to the NetApp as part of a SAN migration. This has been completed successfully on other servers many times in the past with the servers having no issues accomodating both the IBM and NetApp storage at the same time.

Unfortunately when I did it yesterday and performed a rescan via Windows Disk Management all of the already presented LUNs disappeared and they only returned following a reboot of the server.

There were a number of alerts around the time the LUNs were losts which I have detailed below:

Event ID	Event Source	Message
5	sdddsm	Path 7 removed from multipath device 7 by MPIO
17	mpio	\Device\MPIODisk7 is currently in a degraded state. One or more paths have failed though the process is now complete
12	PlugPlayManager	The device 'IBM 2145 SCSI Disk Device' (SCSI\Disk&Ven_IBM&Prod_2145\5&1ef39972&0&000700) disappeared from the system without first being prepared for removal.

It appears as though MPIO believed it lost all of its paths to its existing and new storage however no errors were reported on either the storage systems or the fabric indicating the issue lies with the server.

Has anyone else encountered this issue and if so any ideas where the fault lies?

Thanks in advance,

Steve.

[thomas_glodde]

I see SDD DSM, this is the IBM MPIO driver if i recall correctly. Be sure that the IBM & NetApp MPIO you use are compatible. NetApp support usualy can give you a supported configuration set for your environment, like using MPIO Verison 5.5 and MPIO Verison 3.2 works on Win 2k8 SP2 etc..

Otherwise your system is unsupported and things like these might happen.

[steven_jenner]

Hi Thomas,

Thanks for the response, the server does indeed have the IBM SDDDSM software installed for the IBM LUNs with the NetApp LUNs using the generic Microsoft DSM. This is a supported configuration and works fine on other servers so I believe the issue lies locally with this server and a possible corruption of the IBM SDDDSM software.

In addition I have a number of other systems that are currently configured in this way with no issues. As stated once the data is copied from the IBM storage the IBM LUNs and SDDDSM will be removed I was just curious if others had a similar setup and encountered the same issue.