NDMP Backup Issue
I have two virtual libraries in a NetApp VTL and I'm using NetBackup 6.5.2 on HP-UX 11.11.
The first library has 12 drives, 4 assigned to the backup server, 4 to filer1 and 4 to filer2. These and other filers backed up using NDMP with no issues.
The second library has 10 drives, 4 assigned to the backup server and 3 assigned to each of the filers. Backups from the backup server to it's 4 drives work with no issues but backups to the filer drives using NDMP fail. A tape is loaded but for some reason it's always loaded into a drive that only the other filer has access to. I've checked drive details on the VTL, the filers and on NetBackup and the serial numbers are all correct.
I'm trying to figure out where the issue lies and my interpretation is that the VTL is loading the tape into the incorrect drive for some reason. Filer and VTL cases are open with NetApp but it's slow going.
Does my idea sound plausible? Any feedback is much appreciated.
Comments
Double check the drive serial number and robotic drive nu
You mentioned serial number is right, so perhaps you already checked this, but I recommend double-double checking on it again.
(Well, device configuration wizard should do this properly anyway ... but it's failing now, we better check)
As you might know already, NetBackup keeps robotic drive number and use it to mount tapes.
You have 6 tape drives in your second VTL, so the drives will be of 1 2 3 4 5 6, in your tpconfig -d output - probably with TLD(1), if your TLD(0) is the first VTL.
If the 1st and 2nd and 3rd tape drives in the VTL are hooked up to the seconf filer's nrst2a nrst0a nrst1a, then the tpconfig -d output should reflect the order right, with associated drive path.
filer2:nrst2a being DRIVE = 1 in TLD(1)
filer2:nrst0a being DRIVE = 2 in TLD(1)
filer2:nrst1a being DRIVE = 3 in TLD(1)
And then if the 4th 5th 6th drives are hooked up to the first filer's nrst3a nrst5a nrst 6a, ,then tpconfig -d should be like this.
filer2:nrst2a being DRIVE = 1 in TLD(1)
filer2:nrst0a being DRIVE = 2 in TLD(1)
filer2:nrst1a being DRIVE = 3 in TLD(1)
filer1:nrst3a being DRIVE = 4 in TLD(1)
filer1:nrst5a being DRIVE = 5 in TLD(1)
filer1:nrst6a being DRIVE = 6 in TLD(1)
One good way to verify if this mapping is right or wrong is to run /usr/openv/volmgr/bin/robtest and move "one tape" at a time, from slot 1 to drive 1, then to drive 2, then to drive 3 ... in TLD(1) while running "mt -f status" syntax from the filer's command prompt.
Abe
Anger Management
Thanks Abesama, Great Advice!
Your advice was spot on, the robtest activity compared to the `tpconfig -d` output showed up that the mapping wasn't actually correct for some reason.
I manually changed the drive mappings and lo and behold, problem solved.
Thanks again!
Regards,
Chris
War is peace, freedom is slavery, ignorance is strength....
Would you like to reply?
Login or Register to post your comment.