Symptom:
Top Ten Reasons Tape Drives Go Down
Solution:
10. Infrequent, random failure during "normal" operation.
Diagnosis: Two drive errors occurred within 12 hours. (See:
Technotes 234412, 235839)
Remedies:
A. Increase the number of allowable drive errors by creating a
file named /usr/openv/netbackup/DRIVE_ERROR_THRESHOLD
that contains an override value.
B. Decrease the duration over which drive errors are accumulated
by creating a file named /usr/openv/netbackup/TIME_WINDOW
that contains an override value.
C. Address the cause of the drive error:
* Status 134: V3.4, Patch J080645, during manual backup:
At end of tape, mount request precedes before dismount is
complete -- create /usr/openv/volmgr/DISABLE_RESOURCES_BUSY
on media server
* Unreliable media (near maximum mount count) may need
replacement.
9. drive goes Down following persistent media errors.
Diagnosis: A robot inserts a (physically) damaged tape into the drive.
Remedy: Manual intervention (e.g., remove the adhesive label
that's covering the media access and gumming everything up.
8. drive goes Down with intermittent media errors (Status 85 or 86).
Diagnosis: Tape drive may need cleaning. (See: Technotes 231451,
236697, 201201)
Remedies:
A. Check whether drive hardware has TapeAlert and is configured
for automatic cleaning. (This feature allows a drive to notify
the Media Manager when it needs cleaning. Using this feature
is the Veritas-preferred configuration.)
B. If TapeAlert is not supported (or automatic cleaning is not
selected), check whether Cleaning Frequency is set in the drive
configuration.
C. Find out whether any cleaning tapes are configured, and where
they are located (in the robotic library, in "that other media
server," or in "locked up in Jeremy's office."
7. drive goes Down with Status 85 (media read error) when mounting a
particular tape volume.
Diagnosis: The tape is really a misconfigured cleaning tape.
Remedy:
A. Remove tape from drive (either manually or with robtest).
B. Determine that tape is misconfigured either from its volume ID
or by visual examination.
C. Modify the tape properties (identify it as a cleaning tape.)
6. drive goes Down with Status 85 or 86 for other reasons.
Diagnosis: Further details are available in /var/logs/syslog or
/var/adm/messages.
5. drive remains Down from the time the robot was brought on-line.
Diagnosis: The drive had a tape in it when powered on, and the
robot doesn't know where it belongs.
Remedy: Move tape to an unused slot with robtest and
re-inventory robot.
4. drive goes Down whenever robot loads a tape.
Diagnosis: One or more tape drives are misconfigured. (See Technote #193280)
Remedy:
A. Run "sgscan" and examine the output.
B. Run "tpconfig -d" and compare the results to sgscan.
C. Check for consistency of preceding results, and compare with
/kernel/drv/st.conf
D. Try moving tapes from one drive to the next using "robtest".
Remedy:
* Review current drive definition/properties with tpconfig or
in the bpadm/xbpadm/jnbSA device manager functions.
* Modify as necessary.
3. Hardware problems:
A. Symptom: SCSI Log contains send errors
Diagnosis: Cabling or termination problems (See Technote #234715)
Remedy: Move cables and termination as necessary
B. Symptom: SCSI bus reset
Diagnosis: Defective SCSI controller (See Technote #201880)
Remedy: Replace SCSI adapter
C. Symptom: System error messages from robtest and sgscan
Diagnosis: drive hardware, firmware bugs (See Technote #237227)
Remedy: Upgrade hardware as necessary
Other trouble-shooting tools:
* Check lights on drive
* Other command lines fail
* Errors that appear in various NetBackup logs
* System error logs, such as /var/logs/syslog or /var/adm/messages
(if syslogging is enabled in /etc/syslog.conf)
* /usr/openv/netbackup/db/media/errors
2. Check bptm and daemon logs for clues. Who knows what you'll find in there!?
And the no. 1 reason for tapes to go Down ...
1. drive goes Down at random times.
Diagnosis: Tape drive is linked to a stock index.
Remedy: "/usr/sbin/unlink -f /etc/index.funds" on Media Server.