Re: Migration to EMC Invista.

 
Christof Schoeman
Frequent Advisor

Migration to EMC Invista.

Hi

I have a problem with the migration from EMC Symmetrix to EMC Invista (virtualized) devices. This is the situation.

OS: HP-UX 11i v2
VxVM with DMP for multipathing.

The device in question (an example): c2t2d7
(with alternate path c4t2d7).

Using Invista, 2 additional paths are added, to the same LUN. These are c18t3d0 and c19t3d0.

However, DMP does not recognise these additional paths, while powerpath are well aware of them.

#powermt display nonvirtual dev=c18t3d0
Invista ID=CK200070400727
Logical device ID=60001440B0604774002029600E000145
NonVirtual Symmetrix ID=000xxxxxxxx
NonVirtual device ID=024F
NonVirtual device name=c2t2d7
state=alive; policy=ADaptive; priority=0; queued-IOs=0
==============================================================================
---------------- Host --------------- - Stor - -- I/O Path - -- Stats ---
### HW Path I/O Paths Interf. Mode State Q-IOs Errors
==============================================================================
18 0/0/6/1/0.146.0.1.0.3.0 c18t3d0 0D active alive 0 0
19 0/0/14/1/0.147.0.0.0.3.0 c19t3d0 0C active alive 0 0

Has anyone done this kind of migration before?

Is there a way that I can tell DMP to recognise the additional devices as paths to the same LUN?

Hope you can help.

2 REPLIES 2
Christof Schoeman
Frequent Advisor

Re: Migration to EMC Invista.

Some more information:

I've done ioscan's, insf's, vxdctl enable's, powermt config's and all those nice commands that scans the buses.

The LUNs do not appear to be accessible via the Invista paths:

#diskinfo /dev/rdsk/c18t3d0
SCSI describe of /dev/rdsk/c18t3d0:
vendor: EMC
product id: Invista
type: direct access
size: 0 Kbytes
bytes per sector: 0

#dd if=/dev/rdsk/c18t3d0 of=/dev/null bs=8k
dd read error: No such device or address
0+0 records in
0+0 records out

#vxddladm listjbod
VID PID Opcode Page Code Page Offset SNO length Policy
==========================================================================
...
DGC ALL PIDs 18 131 8 16 Disk
EMC ALL PIDs 18 -1 36 12 Disk

The question is, though, should I be able to address the LUNs via the Invista paths at this stage? If so, where else can I check for connectivity issues?

Eagerly awaiting any response :-)
Christof Schoeman
Frequent Advisor

Re: Migration to EMC Invista.

Yet more information:

The device was set as "Not ready" on the SAN. That was changed to "Ready" and I now have access to the device:

#dd if=/dev/rdsk/c18t3d0 of=/dev/null bs=8k
6520+0 records in
6520+0 records out

Still, DMP does not acknowledge the controller:

#vxdmpadm listctlr all
CTLR-NAME ENCLR-TYPE STATE ENCLR-NAME
=====================================================
c4 EMC ENABLED EMC0
c2 EMC ENABLED EMC0

How can I make VxVM see this device?