Arcada Backupexec -- BEWARE!!!

Thought I'd pass this info along before anybody runs into the same problems I ran into (or worse) . . .

I was testing out the Aracada Backup Exec (version 7.0b for Netware -- the latest version of Backup Exec) and their Solaris 2.4 client, to back up the files on my Sun SparcServer 1000, running Oracle v7.1.4. I ran a few test jobs on files and finally on the test instance of my database, tried a couple of restores, and everything looked great -- backups were fast, everything went smoothly. Finally, I was ready to test a backup of the production database and files -- it started up, and towards the end of the backup, we had to abort the backup job so that we could down the Netware Server that held the tape drives (for a completely unrelated reason). Anyway, we aborted the backup job, everything went fine -- it was in the middle of backing up my redo log files.

Then, I noticed that my redo logs were missing -- not just the one it was backing up, but all the contents of the directory! -- Furthermore, I couldn't restore off of the tape -- the tape's catalog didn't work, and no matter what I did, I couldn't get it to re-index the tape so that I could restore the files. What really drove me mad was that, when I ran the verify tape command, I saw all of the files going by -- but there was no way to get at them!

After all of this, I contacted Arcada (now Seagate Software). I got hold of one of their techs, who put me on hold and checked around -- their answer was, "we've heard of that before" -- no warnings, nothing in the docs, nothing -- yet, they've heard of that before . . . -- they say they've got an Oracle client for Netware, but they haven't developed one for Solaris yet. They say it should work if the databases are shut down, but this can occur when the databases are up, in backup mode or not.

Bottom line -- AVOID ARCADA BACKUPEXEC!!! -- It's bad enough that this can delete your Oracle files at all, but then not to tell anybody about it until it's too late -- I think we have a problem here . . .

Brian J. Hengen                         brianh_at_eia.org
EIA                                                     (703)907-7777