To:	ADSM-L@VM.MARIST.EDU
cc:	 

Subject:	DLT Clean problem



Hi all,

I was wondering if some of you has had the following problem:

--What we have:

*) Storagetek 9740 with V1?????????????????
*) 5 DLT7000 Drives with V115
*) Sun Solaris 2.6 runing TSM V4.1.2.0
*) Autoclean is disabled on the library
*) TSM as "cleaning frequency" on all drives configure to "asneeded".

--What happens:

1) When a drive turns on the "Cleaning Required" led, TSM doesn't receive any 
indication (or does not report it on the actlog ?) and therefore, it doesn't 
clean the drive.

2) In fact, TSM continues to read and/or write to the same tape or to another 
tape loaded afterwards, although the drives's clean led remains on.

3) Then, after a while, the drive reachs a point where it cannot read or write 
anymore to a tape. When this happens, TSM gives the following error:

2001-04-29 19:20:05
ANR8302E I/O error on drive DRIVE_0 (/dev/rmt/7mt)  (OP=READ, CC=422, KEY=03, 
ASC=80, ASCQ=01,
SENSE=F0.00.03.00.00.00.00.16.00.00.8D.28.80.01.00.00.00.00.81.03.00.00.00.00.00.00.00.00.00.,
Description=An undetermined error has occurred).  Refer to Appendix D in the 
'Messages' manual for recommended action.            

Which also appears on solaris syslog: 
  "unix: WARNING: mt@0,0: scsi_cmd=8,errno=110,cc=422,sense=f0  0  3  0  0  0  
0 16  0  0 8d 28 80  1  0  0  0  0 81  3 ^ "

4) TSM now cleans the drive but only after  the ANR8302E message:

2001-04-29 19:20:05      ANR8914I Drive DRIVE_0 (/dev/rmt/7mt) in library 
STK9740_2 needs to be cleaned.
2001-04-29 19:23:37      ANR8906I Cleaning drive DRIVE_0 (/dev/rmt/7mt) in 
library STK9740_2.
2001-04-29 19:34:13      ANR8908I CLEAN DRIVE for drive DRIVE_0 (/dev/rmt/7mt) 
in library STK9740_2 completed successfully.                


--The problem:

Due to the ANR8302E, any operation that was using that tape/drive fails. If the 
operation that failed had been, for example, a reclamation process, there 
shouldn't be a problem because we can allways restart a reclamation process.

However, if the operation that abort had been, for example, a TDPOracle backup 
of a huge Database, then this is a problem because the backup will fail and it 
will have to start all over again.

There is also another problem. Due to the several clean erros, some tapes are 
now unreadable and we already lost some. I managed to recover some with 
"restore volume" command but, there were other that I couldn't recover because 
the readings kept failing, precisely when they were being copied to a  copy 
pool.


--The simple question:

Why doesn't TSM cleans the drive right after the drive's clean led turns ON ?


Thank you very much,
Justino Miguel Costa
Edinfor - Unix Team
justino.costa@edinfor.edp.pt