This is a strange case I encounter with a customer:

[ NW 7.5.1 on a HPUX, hardware maintenance updated all drivers/opcodes of the os, library and tape drives ]

the daemon.raw output:

38758 04/18/10 12:36:46  2 0 0 1 2309 0 mercator nsrd media critical: TapeAlert flag "Write failure" has been posted for /dev/rmt/c8t15d0BESTnb
38758 04/18/10 12:36:46  2 0 0 1 2309 0 mercator nsrd media warning: TapeAlert flag "Hard error" has been posted for /dev/rmt/c8t15d0BESTnb
38758 04/18/10 12:36:46  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb writing: I/O error, at file 70 record 13338
42506 04/18/10 12:36:46  2 0 0 1 2309 0 mercator nsrd media notice: sdlt600 tape 001005 on /dev/rmt/c8t15d0BESTnb is full
42506 04/18/10 12:36:46  2 0 0 1 2309 0 mercator nsrd media notice: sdlt600 tape 001005 used 144 GB of 160 GB capacity
42506 04/18/10 12:36:46  2 0 0 1 2309 0 mercator nsrd media info: WORM capable DLTWORM capable for device /dev/rmt/c8t15d0BESTnb have been set
34353 04/18/10 12:36:46  2 0 0 1 5332 0 mercator nsrmmd tape_bsf bsf failed: drive status is The end of data was reached
34353 04/18/10 12:36:46  2 0 0 1 5332 0 mercator nsrmmd tape_bsf bsf failed: drive status is The end of data was reached
38758 04/18/10 12:36:46  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb moving: tape_bsf bsf failed: drive status is The end of data was reached
34353 04/18/10 12:36:47  2 0 0 1 5332 0 mercator nsrmmd tape_bsf bsf failed: drive status is The end of data was reached
34353 04/18/10 12:36:47  2 0 0 1 5332 0 mercator nsrmmd tape_bsf bsf failed: drive status is The end of data was reached
38758 04/18/10 12:36:47  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb moving: tape_bsf bsf failed: drive status is The end of data was reached
 
42506 04/18/10 14:30:08  2 0 0 1 2309 0 mercator nsrd media info: WORM capable DLTWORM capable for device /dev/rmt/c8t15d0BESTnb have been set
38758 04/18/10 14:30:08  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb reading: I/O error
 
42506 04/18/10 14:50:28  2 0 0 1 2309 0 mercator nsrd media info: WORM capable DLTWORM capable for device /dev/rmt/c8t15d0BESTnb have been set
38758 04/18/10 14:50:28  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb reading: I/O error
 
42506 04/18/10 15:10:48  2 0 0 1 2309 0 mercator nsrd media info: WORM capable DLTWORM capable for device /dev/rmt/c8t15d0BESTnb have been set
38758 04/18/10 15:10:48  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb reading: I/O error
 
42506 04/18/10 15:31:08  2 0 0 1 2309 0 mercator nsrd media info: WORM capable DLTWORM capable for device /dev/rmt/c8t15d0BESTnb have been set
38758 04/18/10 15:31:08  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb reading: I/O error
 
42506 04/18/10 15:51:28  2 0 0 1 2309 0 mercator nsrd media info: WORM capable DLTWORM capable for device /dev/rmt/c8t15d0BESTnb have been set
38758 04/18/10 15:51:28  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb reading: I/O error
 
42506 04/18/10 16:06:51  2 0 0 1 2309 0 mercator nsrd media info: WORM capable DLTWORM capable for device /dev/rmt/c8t15d0BESTnb have been set
 
38758 04/18/10 16:09:26  2 0 0 1 2309 0 mercator nsrd media warning: /dev/rmt/c8t15d0BESTnb reading: I/O error
42506 04/18/10 16:09:26  2 0 0 1 2309 0 mercator nsrd media notice: Volume "001005" on device "/dev/rmt/c8t15d0BESTnb": Cannot decode block. Verify the device configuration. Tape positioning by record is disabled.

Before we start the discussion, the problem mentioned is at random in time and not always on the same tape drive. 

The policy is: if a tape drive or a tape reports 3 times the same problem on a physical level, the device or tape is replaced. Most of the tape drives are relative new.

The first two lines concerning the TapeAlert indicates that there are problems on tape drive hardware level (the connection or the tape drive itself) . If this was the case we would see this message returning in the daily logs, which is not the case ...

Then the I/O writing error emerges and finishes the save stream to the tape and marks the tape as full. At this time, NetWorker should unmount the tape and mount another one continuing the backup. Now for some reason it tries to read-in the tape again (to get the marker of the backup stream?) and the time wasting starts. First we see (last message at 12:36h and the next at 14:30h) the tape device is hold for nearly two hours; then NW retries several times to read in the tape again (interval 20 minutes, 6 retries) . How can we decrease this time wasting? The last message "Cannot decode block. etc" is in the NW error message guide. But the solutions provided or the things to watch are satisfied.

Even though we have this error, backups are running fine. NetWorker does its job and we are operational.

Any ideas? Things to discuss?