PROBLEM:
The operation failed because of a protection agent failure. (ID 998 Details:
An existing connection was forcibly closed by the remote host (0x80072746))
SYMPTOMS:
Had this problem occouring when trying backup to tape.
The majority of the tape backup jobs worked fine, then on two of the disks on two of the servers it kept failing
interestingly failing at about the same place. In our case at 3.5GB on one disk and 700MB on the other.
After getting Dell to replace the TL2000, upgraded firmware for the library and the drive, used the latest drivers from DEll and
from IBM as its a IBM rebranded library and drive still the same problem.
Patched up all the hotfixes i could find at the time, SP1 of DPM 2007 too. Still no joy.
This doesn’t make sense as the backup is from the DPM replica to tape, it is NOT connecting to the original host.
SOLUTION/WORKAROUND:
I was with our Anti-virus guy (Kaspersky) and was discussing the problem, although i’d given him the list of the DPM volume
mount points used to exclude from the scan the AV software still scanned the replica and volume shadow copies on the DPM server.
I retried the failed job and waiting, about 5 seconds before the tape job failed the AV detected 3 virsues on the disk and tried
to delete them, as its in a volume mount point, it couldn’t and therefore blocked access, 10 seconds later the tape backup job reported failed.
So that was it the AV was causing the problem. Removed the AV software (to test) and reran the job went through no problems.
Now we have working backups the challenge is to get them to work with the AV installed (Kaspersky in our case)