False use of DLT cleaning tape?

[A. Daniel King_1]

Hi, folks.

Is there a way to determine the usage of a DLT cleaning tape - from the tape itself? Are there visual queues to determine if a cleaning tape has been used - or used up?

The situation is this:

Just recently, one of our tape silos started using a cleaning cartridge every three days - down from about two weeks previously. There was no corresponding increase in usage. I wonder if the software/firmware is reporting false statistics or if it is using the cleaning tapes excessively. Can you tell this from a tape?

Any thoughts appreciated.

[S.K. Chan]

I don't think so. Typically a DLT cleaning cartridge takes up to 20 cleaning cycles. Did you had "autoclean" turned on from the tape silo side ? If yes you may want to turn it off and let you backup application manage the cleaning instead.

[A. Clay Stephenson]

No. I suspect that you have a drive going bad (or possibly a medium failing). Most tape libraries have commands to make them dump statistics; you can also use STM to examine the read and write errors for each drive. If you see significantly more corrected read or write errors on one drive than on the others then that is your boy and you should have the drive replaced.

[A. Daniel King_1]

Thanks for the prompt replies. I've really got to learn something about stm:

- I find that I can use it as myself instead of root (good).
- I don't know if I can run some of the options while the tape drives are in use - do the dignostic tests require the use of a scratch tape? Can some tests be run without disconnecting the application using the drives?

Any additional information appreciated.

[A. Clay Stephenson · Accepted solution]

You don't even have to run the diagnostics and no tape needs to be mounted. I'll talk you through the GUI version of STM because it is so easy to use (if you have an X display).

1) xstm
2) Locate your DLT drive in question and select it via a left mouse click.
3) Tools->Information->Run

After a few seconds a dilog box will appear will the current statistics stored in the drive. The main values you are interested in are 'Read Errors Corrected', 'Write Errors Corrected', 'Read Errors Uncorrected' and 'Write Errors Uncorrected'. All of these should be VERY NEARLY zero on a good drive. All of mine on all drives are currectly 0 but on a bad drive you will see non-zero vales. That is then your suspect drive.

You can also run the Diagnostics which does require a scratch tape.

[A. Daniel King_1]

The GUI is fairly easy.

However, next to the device listing, I get "Information Incomplete", and the Information Tool log contains only header, date & HW path information.

Other devices do not seem to do this (i.e., fiber-channel cards) - these seem to give actual statistics.

I'll search on this, but your input is greatly appreciated.