Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Chris Miller on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Arcserv 2000 SP4 merge failures :-(

Status
Not open for further replies.

gwichman

IS-IT--Management
Sep 10, 2002
199
US
Well so far CA's support has not been very helpful. So lets try a post here. I have hundreds of tapes we did not create and i'm attempting to merge them to see whats on them. About only 30% merge successfully.. The others error out at some point with one of 2 things happening:

One:

E3801 Invalid session header signature. (SIG=5555aaaa)

With this error it typically occurs on the 1st tape of the set of tapes (sequence 1). the last tape in the sequence i can usually merge fine if i do it individually. The middle tapes usually throw error Two listed below for some session #.

Two:

This is more common among the two. At least 50% of the errors are this one:
W3885 Unable to position media catalog file.
E3841 Unable to find session number (MEDIA=<tape label>, SESSION=<some session #>, EC=EARLY WARNING)

Usually the 1st tape merges fine with this error and one of the subsequent sequences fails witht he above error. When i merge the last tape in the set, it merges fine but all the middle ones throw the above error usually with the same session number.

I'm at a loss as to what to do about it. CA wants me to investigate on whether the devices are compatible (the ones that created the tapes vs ours which we're using to read them). Both are DLT8000's which i understand Quantum makes all those drives anyhow. So perhaps more of a firmware issue like others have speculated on below. Our firmware is relatively new (232 in one Adic single DLT8000 drive, 250 on our Quantum tape library DLT8000 drives). I was told the device that created the tapes was an ATL 1000 auto changer with 3 DLT drives but can't confirm it personally nor do i know what firmware the drives were.

Next step beyond posting here is to talk to Quantum and see what insight they have as well as continue to poke CA's unhelpful support along. Appreciate any feedback
 
The drives could be the same but you may have to check the blocksize those tape were written to. If you current drive is set to read 64K blocksize, and the other is set to 16K block size, then this may be the issue. Also are you sure the other tapes were written with ARCserve?

Thanks,

g
 
According to Arcserve's website, i read someplace that with SP4 (and 3 i believe) the block size is automatically deteremined. So wouldn't Arcserv simply determine the blocksize when i inserted the tape?

Is there some way to determine the blocksize of a tape created 2 years ago that perhaps nobody knows what the blocksize was?

One problem i'll admit i'm having is getting a straight answer as to what created the tapes. I was first told Arcserv. Later Netbackup. Netbackup clearly can't read them at all as we've tried. So that leaves Arcserv. Trouble is they had Arcserv running on both Novell and NT. We are trying to read them with Arcserv&NT with some success. Would an Arcserv/NT server be able to read Arcserv/Novell created tapes?

I tried reading some of the tapes with Veritas Backup Exec and it was able to inventory them (read the tape label) but not catalog them (same as merge on arcserv). The error log indicated the tapes were arcserv compress format.

We have a conference call tomorrow to iron out some of these issues and get better answers. It seems the questions to ask (in some cases, once again) are:

-device firmware level of tape drive that created tapes
-OS & version of arcserv (or other product) that created tapes
-block size if possible (i'm skeptical they'll know this)

Any additional thoughts certainly appreciated
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top