Back a while ago, I posted a similar discussion about the trials and
tribulations of tape drive (TS1120/TS1130) replacement and the work it
caused when the serial number changed (rebooting and reconfiguring all TSM
servers).  I was enlightened that the new drives serial number could be
reset to the old drives serial number and the issue of having to reboot
all 7-TSM servers because of a drive replacement, was, I thought,
resolved.

Well, I haven't had a need to drive replacement since then........until
now.

What I discovered was that even with the replacement drive co-opting the
serial number of the failing drive, I still have to reboot all of my
servers.

Yes, I made sure to update the new drives serial number BEFORE plugging in
the fibre cable.

Took the library and drives offline in TSM

Yet, as soon as the failing drive was unplugged (or at least it seems that
way), the /dev/IBMtape? device becomes unavailable and no amount of
stopping/restarting the lin_tape processes will make the drive accessible.
Reboot is required.  Any attempt to open the drive via IBMtapeutil fails
(Operation failed with errno 5: Input/output error).

So, what are we doing wrong?  How do you handle tape drive
repairs/replacement?

To recap my environment:

IBM3494 with TS1120 and TS1130 drives connected via fibre
CISCO SAN switches
Single path to each drive (not enough switch ports)
RHEL4 and RHEL5 servers
My SAN guy says we are not mapping drives by WWN - pools/groups I think
(sorry - I don't know the terminology) - all tape drives and servers that
use them are separate from disk connections
Zoltan Forray
TSM Software & Hardware Administrator
Virginia Commonwealth University
UCC/Office of Technology Services
zforray AT vcu DOT edu - 804-828-4807
Don't be a phishing victim - VCU and other reputable organizations will
never use email to request that you reply with your password, social
security number or confidential personal information. For more details
visit http://infosecurity.vcu.edu/phishing.html