Automatic cleaning fails, drive goes offline

[MPieper]

Hi,

since i upgraded my Backup Exec 2010 installation to ServicePack 1, i have an issue with the Auto Cleaning in Backup Exec.

When Backup Exec tries to AutoClean a drive, i get a message that no cleaning Tape is found, and the affected tape drive goes offline. I have to set the tape back to online manually. The Job using that drive is queued while the drive is offline, but continues normally after i set it to online.

I have a cleaning Slot defined, and there's a working cleaning tape in that slot. When i do a manual cleaning it works fine.

I have a DELL ML6000 autoloader library with 8 drives, no cleaning slot is designated in the library. All firmware is up to date. I use the latest manufacturer drivers for the drives. Backup Exec 2010 is also up to date, a migration to 2010 R2 is not planned at the moment.

[pkh]

I am confused.  In the 3rd paragraph, you said that you have a cleaning slot defined and in the last paragraph you said that no cleaning slot is designated.  Which is which?

You got to define a cleaning slot for the cleaning job to work.

Also, if you are using barcode labels, did the label for the cleaning tape begin with CLN?

[MPieper]

Sorry if i was not clear enough.

The library itself has also an autocleaning feature. I do not use this, only the autocleaning from Backup Exec.

And yes, i use Barcodes, and yes, the barcode starts with CLN. The Media Type was correctly identified a s CLN, also the description says "(Cleaning Media)"

[pkh]

In BE, you must designate the slot that hold the cleaning tape as a cleaning slot.  Check the cleaning slot option when you right-click on that slot.

[MPieper]

The slot is designated as cleaning slot. When i right click on a tape drive and select "Clean" this works normally. The "Since Last Cleaning" statistics of the drive reset, and the count on the cleaning tape increases by one.

[pkh]

What is the result if you schedule a cleaning job for later?

[MPieper]

A scheduled cleaning job ran fine.

The problem seems to occur only when a drive requests cleaning while a job is running on that drive. Whether this is a routine cleaning or a required cleaning does not seem to matter. Before the update this worked fine. The job paused, the automatic cleaning ran, and the job continued afterwards automatically. Now the job hangs on queued, since the drive goes offline after the failed cleaning.

[riva11]

Did you try to define a clean slot on library and run cleaning from library ? This can be also helpful to understand if there is an issue on clean media tape.

[MPieper]

To define a cleaning slot in the library itself, i would have to recreate its partitions. This is not really feasible at the moment since i have currently still two backup jobs running.

When i do a manual cleaning from the library webinterface this works fine, too, but Backup Exec does not notice it when i clean via the webinterface. I see that it works, only because Backup Exec does not prompt for cleaning again. 

[MPieper]

Here is an update:

One of my running jobs with this issue just finished. I got the alert again that the drive needs cleaning. And this time, the automatic cleaning worked!

So my problem seems to occur only when a tape requests cleaning while the job is not finished yet.

I would turn the autocleaning off, but from time to time one of my drives reports a required cleaning and stops until it is cleaned.

Is it possible to configure the automatic cleaning to only try to clean the drive when the drive reports a required cleaning, as opposed to the routine cleaning, which could be ignored over the weekend?

[Larry_Fine]

re: I would turn the autocleaning off, but from time to time one of my drives reports a required cleaning and stops until it is cleaned.

How do you turn autocleaning off?

I have never seen a tape drive stop operating due to it desiring to be cleaned.  Usually the drives just keep sending TapeAlerts.  And I have never seen cleaning occur in the middle of a job, I thought the cleaning request was always queued and ran after the job completed.

[pkh]

I take it that your tape drive kept asking to be cleaned.  From my experience, this is a sign of failure.  I had a similar situation about a year ago.  My tape drive kept asking to be cleaned and my tapes start to have a lot of soft errors.  Luckily, my tape drive is still under warranty, so I requested for a change and the cleaning problem went away.

[MPieper]

As soon as i clean the drive requesting cleaning, wether i clean trough Backup Exec or the webinterface of the library, it stops to ask. The cleaning LED on the drive itself turns off, too. Normally, it asks again after about 200 in use hours, which is normal behaviour according to DELL.

Sometimes, despite the regular cleaning, one of my drives (not always the same) asks for a required cleaning and stops entirely until it is cleaned. This is why i started the cleaning through Backup Exec in the first place.

One of my drives shows 27 hard write errors and 3 soft read errors since 30th April 2010. The drive itself is about 2 1/2 years old and still under warranty. On a side note: the statistics for that drive seem borky... it shows 218,104,243.1 Terabytes of total data written. The statistics for my other 7 drives seem ok, with either none or only a few soft errors and a reasonable amount of data written.

I will now remove the cleaning tape from Backup Exec, to see if the drive still goes offline even when no cleaning is attempted. The next drive should request cleaning during our next weekend backup. I will let you know how this turns out.

[MPieper]

After i removed the cleaning Tape from the library, and reconfigured the cleaning slot as normal storage slot, still get the information alert that Backup Exec attempts to automatically clean the drive, everytime the drive requests cleaning. Obviously, afterwards i get the alert that no cleaning media is found, since i removed it. But now the drive does not go offline.

The cleaning request occurs everytime one item in the selection list is finished, or when the job rolls over to the next medium. Since the job doesn't stop anymore i can live with those warnings.

I would prefer to keep a cleaning media within Backup Exec, but since it seems that my error is somewhat obscure, i most likely have to live with the problem ontil hopefully some future hotfix fixes it...