Backup Exec 8.6 --- Creates unrecognizable media???

[NCCAdmin]

Hey Everyone,

I have a pressing issue to deal with concerning Veritas Backup Exec producing "Unrecognized Media".

Here are the details. I'm backing up roughly 20 servers, including a few Citrix Application Server, an Exchange mail server, a financial server, a file server, a development server, and a few domain controllers.

I have have installed the Veritas Backup Exec 8.6 Windows 2K agent on every server (...every server is Windows 2K SP4). The backup server is currently also operating as our third DC (for redundancy purposes) and includes a clean install of Backup-Exec 8.6 and the Exchange Agent. SQL agent is not needed, because I run a script to backup SQL data to zip files, which is backed up by BE.

For every server, I have created a separate backup job. Every job is set the same: FULL, with open file support (with a lock), hardware compression enabled, etc.

Each job is also set to "overwrite if NO appendable media exists".

Currently, I have configured the backup server to use the Exabyte Mammoth EZ17 Tape Loader unit, which includes a built-in Mammoth-2 drive. The unit has been configured correctly, and Backup Exec seems to have no problems utilizing the device (reading tapes, writing tapes, inventory, etc.).

So, here is the problem:

Each of the backup jobs is set to run nightly, one after the other. Each job appends to where the last job left off. Each job, other than the file server backup, is relatively small (...each is less than 8 GB, except for the file server, which is 130 GB). Everything seems to run fine, except, when the file server backup is complete, it renders the third tape in the rotation as "unrecognized media"... ...the job that immediately follows, is forced to use the next available tape in the rotation. I have checked the tapes that become unrecognized, and they are fine physically. I have exchanged various tapes, and have replaced older tapes with brand new ones, and the problem still happens. I have checked the logs, and nothing weird is reported, other than the occasional "corrupt file" being backed up (ie. the occasional Outlook "archive.pst" file).

Is there anything that I'm doing wrong? Each Mammoth tape can hold 60 GB uncompressed. All other servers in our NOC (other than the bricks mailbox backup which is scheduled to run after the file server backup) is easily able to make it on to one tape. The file server backup then appends to the first tape, writes through the second, and completes on the third. However, when it finishes, the third tape is rendered "Unrecognizable media", until I re-label it. This is a serious problem because ALL of our backups are being corrupted, as the third tape is unavailable when we need to perform a restore.


Please help. I'll try anything.

 

[justin2000]

Just a question, if you are running the jobs one after another, and after the first job, the jobs will append to the last tape used - and you say your backups are about 8gb except for 1 server - and the tapes are 60gb - and you say after it does the file server backup - it uses a new tape for the next job.. how can ALL your backups be corrupted?
So what you are saying your can't restore any data at all?
Has this just started happening?
Have you swapped out the tape drive?

Also do you run a verify after the jobs finishes?

Are you running 8.6 3878, with any other patches you may need for your environment.

Have you updated tape drivers to the latest set?

 

[NCCAdmin]

"Just a question, if you are running the jobs one after another, and after the first job, the jobs will append to the last tape used - and you say your backups are about 8gb except for 1 server - and the tapes are 60gb - and you say after it does the file server backup - it uses a new tape for the next job.. how can ALL your backups be corrupted?"

OK, to answer your question, usually, when performing a restore, I get asked to insert the last tape of the backup set. Since the third tape cannot be read, the entire backup becomes useless. In other words, the entire 130 GB file backup is unaccessible. Any suggestions?


"So what you are saying your can't restore any data at all?
Has this just started happening?"

It only started happening when I decided to make a separate backup job for each server. Before this, I had set the backups to work as follows: One job was configured to backup ALL servers, including Exchange, and was labelled as "All Servers". Another job was configured to backup just the file server, called "File Backup". This setup did not seem to corrupt any tapes. This is why I'm beginning to think that it must be something that I've configured wrong, but for the life of me, I can't figure out what.


"Have you swapped out the tape drive?"

This is not the issue. It is a brand new tape drive, and this problem was happening on the old one as well. Both are Mammoth-2 tape drives, but the new one is part of the EZ17 Autoloader system.


"Also do you run a verify after the jobs finishes?"

No. I do not run verify jobs, because if a single file in the entire verify is corrupt, Backup Exec does not allow me to restore any files beyond that point.

"Are you running 8.6 3878, with any other patches you may need for your environment."

I believe so, but let me double check the 3878 code.



"Have you updated tape drivers to the latest set?"

I'm using the latest Windows 2K drivers for the EZ17 Autoloader... ...not Veritas' built-in drivers.

 

[justin2000]

NCCAdmin

have you seen this technote?

http://seer.support.veritas.com/docs/239808.htm

Check it out it may resolve your problem.

I would also suggest to use the Veritas tape drivers as well, as they have been tested with the device..but.. with limitations from the technote above.

One thing I was thinking which could also be the problem - have you set the compression on all the jobs to be the same? Check this.. I don't know what exactly would happen if you set one job to use hardware compression, then the next job that runs to not use compression, or use software compression.

If you setup a backup job to backup all the servers, does the issue still exist?