Tape Libraries and Drives
cancel
Showing results for 
Search instead for 
Did you mean: 

Regular backup failures due to MSL5000 issues

Mark Bainter
Advisor

Regular backup failures due to MSL5000 issues

Hardware:
Compaq MSL5000 w/two drives
Compaq MDR
Compaq Arbitrated Loop Switch
Compaq DL580
Software:
Veritas BackupExec
Windows 2k, SP3

Misc related info:
All of the hardware in question is current firmware wise. The controller card for the robotic arm has been replaced twice in two years.

Problem:
On a regular basis, veritas hangs at some point in the backup. Usually surrounding loading or unloading the tape.

Sometimes I can just cancel the job, restart services, etc and it'll come back. Sometimes rebooting will do it. I can always disable the devices in backup exec and re-enable them to get them to come back in backupexec, but they go back offline as soon as I try to do anything with them. Most of the time I have to clear the fcscsimap on the MDR, and then reboot the tape library, and then sometimes the server as well, to get everything to come back. Two days later (at most) the problem will be back again.

I don't see any weird errors in the tracelogs from the MSL5000. No indications of hardware failure there.

Anyone have any ideas on what I can do to make this run more reliably? Constantly missing backups is not something I'm comfortable with. I'm terribly disappointed in the quality of these products.

1 REPLY
Eugeny Brychkov
Honored Contributor

Re: Regular backup failures due to MSL5000 issues

I would begin with:
1. moving to fabric or zoning library and host ports into separate zone. Why? Because a) in FC-AL loop if any device on the loop will issue reset command all the loop devices may got reset; b) other hosts may be polling device confusing backup device and causing backup to fail;
2. checking any software which may be polling devices and causing problems;
3. drivers setup - latest, driver parameters, etc. If you're using CPQ FC HBA make sure you're using CPQ driver. Check http://www.emulex.com/ts/docfc/w2kmini/5481a4/tocconfigure.htm (I believe CPQ driver has the same parameters). Pay attention to 'ResetTPRLO' and 'EmulexOption' parameters
Eugeny