John,

the answer is that someone have put a SCSI adapter in the box and Windoze
being "so smart" reordered ALL the SCSI/FC adapters. SCSI adapters get
numbered before FC ones thus your QLogic get higher number (6, having 5 in
the past). The other possibility is that there was second SAN switch/plane
installed. Again additional adapter but insertion of FC adapter is
somewhat 50/50% chance - PCI slot numbers are considered which one is
getting lower number. If you are unlucky - new inserted FC adapter gets
HBA #5 and your one moves to HBA #6.
If it was AIX - new adapter will get next *available* device name (scsiX
or fcsY+fscsiY) preserving existing device names.
Keep in mind also another thing to be aware - the SDG module in 3583. If
someone makes SCSI channel to FC port zoning, changing number of visible
drives (for example willing to split each device to separate FC ports)
this may affect \\.\Tape0. And \\.\Tape1 can become \\.\Tape0 leading to
element number mismatch - TSM will refuse to work with the drive.

Zlatko Krastev
IT Consultant




Please respond to "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
Sent by:        "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
To:     ADSM-L AT VM.MARIST DOT EDU
cc:

Subject:        3583 with SDG, Windows 2000 and Qlogic 2200 adapters

Hi all

Here's the scenario:

TSM 4.2.1.15 Server on Win2K
3583 with San Data Gateway
2109 Switch

Implemented the solution only to be called back with the library not
responding to TSM. After a closer inspection I notice that the library has
now been renamed by Windows from lb2.1.0.5 to lb3.1.0.6.

So, how is this possible. There have been no hardware address changes made
and I made sure to name the libary and the drives independently of the
windows generic naming ie

lb2.1.0.5 was configured as LTOLIB
\\.\Tape0 was configured as DRIVE1
\\.\Tape1 was configured as DRIVE2

I have read in the Qlogic literature that persistent naming is not
supported which means the binding needs to happen elsewhere. I cant zone
the switch to tightly because I will be configuring SAN agents next and
that means all the SAN attached hosts need access to the library.

Has anyone one out there seen this happening in a SAN environment and how
I can make it go away?


Any help is much appreciated

Rgds
John