BE 10d (10.1R5629) Disk to Tape backup fails at end of first tape

[Ceechtay]

Hi all,

I really have a few questions, but the first is the problem in th emessage title. Weekly, we first do a full backup of an IBM xSeries 236 server to an attached eSata 1TB drive(so Disk to disk is successful). That backup has always worked, although very slowly (only getting about 10GB/hour speed, but this does involve backing up Exchange). We then simply want to do a duplicate copy of that full backup from the attached HD to the tape drive(Disk to tape fails). The tape drive is a VXA2 SCSI 320 drive. When we attempt the tape backup, it fails at the end of the first tape with an error of:

e00084f2 - An attempt to write database end of media has failed.

We have retensioned and cleaned the drive heads, but still fail every time. We do daily differentials that work everytime with disk to tape, but those backups involve only one tape. (The disk to tape works at about 25GB/hour speed)

Any ideas outside of trying brand new tapes?

Any ideas about the relatively slow speeds? The tape drive and the SCSI controller both have out-of-date controllers or firmware which need to be updated. Obviously, this will be first thing we try.

Thanks for your help.

[Khue]

How are you doing the back up to disk?

I have this same issue when I have 2 resources writing to the same media at the same time. For instance, job a and job b both write to the "Disk Backups" folder. The disk backups folder is large, say 500 gigs. The 2 jobs are extremely slow. Occasionally the bkf files become "overwriteable" due to the extreme length of the 2 jobs. Job a is 240 gigs job b is 247 gigs. It seems to me that at some point, job a or job b overwrites a piece of media before it gets to tape. 

My suggestion is to create another b2d folder and break out the resources individually. Place the exchange in an b2d exchange folder on the esata drive by itself and place the file backups in their own b2d folder. So basically break out the 1 job and make 2.

As far as the speed goes, it sounds like you are getting about 427mb/min which isn't terrible. Really it depends what type of information your backing up. If you are doing Exchange using GRT, b2d operations should be fairly quick (mine are typically about 1500mb/min). If you are doing file based backups, depending on the size of the files and the file size variance across the resource I would expect speeds from 10mb/min all the way up to 800mb/min. For example, if you had 100 gigs of 10kb files, I would expect that to clock in at 10-100mb/min. If you had like 100 gigs of 20 gig files, I would expect that to come in around 800+mb/min. 

Just my 2 cents.

[Ceechtay]

Thanks for the reply.

Hopefully I am understanding you correctly.

We are running just one job at a time. Nightly, we run one disk to disk job and then the disk to tape job starts after that D2D is finished.

The overwrite protection is 2 days and the jobs don't take that long.

We could easily break it into two and that will be the fix for now. And thanks for the advice on the speed. The Exchange backup is file based and involves a lot of small files.

Tanks again.