[ofa-general] SRP target sporadic behaviour with Solaris, VMware

[Vu Pham]

> 
> I've recently set up the SRP target module on Linux (2.6.22).
> 

How did you set up your target (scsi pass-thru, blockio, 
fileio, nullio)?



> 
> On VMware:
> - I had to reboot my new VMware ESX server a few times before it found 
> my 500GB target.

What is your fabric configuration (switch, which SM, which 
HCA, fw version)?
Do you see any *weird* srp message in vmware /var/log/vmkernel?


> - VMWare completely rejects a target if it doesn't have a partition 
> table - I ran parted on Linux and then VMWare was OK

It does not make sense. Once again I need to see how you 
setup and expose your luns on srp target.


> - Also, the messages in VMWare gave me the impression it would clobber 
> the whole volume, rather than just a single partition - so to avoid the 
> possibility of losing my other partitions, I made a special target 
> representing the intended partition rather than the entire volume.  Now 
> I have a VMware partition table nested within a partition.
> - VMware only seems to show one target at a time - I had created a few 
> test targets, but I could only see one of them.  Is this what other 
> people see?  ibsrpdm on the other Linux hosts shows all the targets.
> 
> Any suggestions?
> 

If you have several srp targets running blockio or fileio 
modes and export the same block device names (ex: all srp 
targets export /dev/cciss/c1d0 device ie. echo "open vdisk0 
/dev/cciss/c1d0 BLOCKIO" >  /proc/scsi_tgt/vdisk/vdisk) then 
they end up to have same lun UID. Vmware will *correctly* 
see single lun with multipaths to it instead of multiple 
luns of multiple targets. To correct this you need to load 
scst_vdisk with different scsi_vdisk_ID for different srp 
targets exporting the same device names in the fabric

srp target 1: modprobe scsi_vdisk scst_vdisk_ID=1
srp target 2: modprobe scsi_vdisk scst_vdisk_ID=2

If you already create datastor on this lun you need to 
destroy it; otherwise, vmware won't see the datastor 
(because of different scst_vdisk_ID will generate different 
lun UID and vmware's datastor relies mainly on lun UID)

-vu