ARCSERVE 2000 End of backup failed to login to DB error

[mleggatt]

We are running ARCSERVE 2000 on a Windows NT4 server which has been running for some time without errors. Now each morning the Database Engine service is either not running or we are unable to view any records in the database (from DB manager). Looking in the ARCSERVE log it seems to fail to register the media and complete the backup and then fail to write to the records to the database. When I look in the ..\ComputerAssociates\ARCserve\Temp directory there are .CAT files. From what I have read I believe there are several databases ARCSERVE uses to keep it's records in e.g. TAPE, SESSIONS, and one of these is corrupt. When ARCSERVE tries to write from these corrupt records the DB engine stops/crashes and the temporary .CAT file is left in the temp directory. My question is how to fix this corruption, from what I have read there is one of two ways:

1. Fix the existing database using the VLDB tools Ddbchain, keybuild, db check and dbfix?

2. Reinitalise the corrupt database.

I am reluctant to use the reinitalise option as 1. I am not sure which database to reinitalise and 2. I am concerned I will lose the backup records and not be able to restore them.

With option 1 I am not really sure which utilities I need to run. I did read the FAQ posting here which suggests:

***********************************************************
Dbfix can be run to repair corruption, however running dchain and then keybuild on a database will do a much better job and is almost always successful in repairing any corruption.
***********************************************************

I did try running the DBFIX utility but it ran for hours and the next nights backup had to be run so it was stopped. After reading this forum it seems these DB utilities can take hours to run approx 6hrs per Gb (Our main arcserve DB is 3Gb).

I was thinking of using a scheduled batch file to schedule the DBCHAIN and KEYBUILD to run Friday night after the back up leaving it the whole weekend to work and see what happens. I also read that some of these utilities should be run on a regular basis to avoid corruption and improve performance, is this the case?

If anyone has any suggestion or anwser to the above I would appreciate it.

Thanks
Martin


0050126 222019 2161 3,516.68 MB Written to Media.
20050126 222019 2161 Elapsed Time: 6m 55s
20050126 222019 2161 Average Throughput: 508.43 MB/min
20050126 222019 2161 ARCServe Database Backed Up
20050126 222019 2161 E4115 Unable to register media. (EC=Error Code = -2009)
20050126 222019 2161 W3502 Unable to add session to database. (EC=-2009)
20050126 222021 2161 E4107 Unable to update job record. (EC=Error Code = -2009)
20050126 222021 2161 E3530 Failed to login to database.(EC=-2005)
20050126 222021 2161 N5936 The following media can be used in device group GROUP0 for next job on date 1/27/05:
20050126 222021 2161 N5941 1. <BLANK>
20050126 222053 [CAT] Failed to login to ARCserveIT database.(EC=-2005)
20050126 222053 [CAT] Failed to login to ARCserveIT database.(EC=-2005)
.
.
.
.
20050126 222140 [CAT] Failed to login to ARCserveIT database.(EC=-2005)
20050126 222201 2161 ** Summary for Job **
20050126 222201 2161 29 Sessions.
20050126 222201 2161 138,916 Directories 731,390 Files (34,053.78 MB) Backed Up to Media.
20050126 222201 2161 77 Directories/Files Skipped.
20050126 222201 2161 35,449.81 MB Written to Media.
20050126 222201 2161 Elapsed Time: 2h 38m 27s
20050126 222201 2161 Average Throughput: 223.71 MB/min
20050126 222201 2161 Number of Errors/Warnings: 1/4
20050126 222201 2161 Backup Operation Incomplete.
20050126 222202 2161 Reschedule Backup Job for 1/27/05 at 7:30 PM.