11-03-2010 01:20 PM
I've read the entries about changing the avalable_media script and I did that. But I still have quite a few tapes showing up as DBBACKUP. When I look at that status of the DBBACKUP tapes they are all 0x0. I thought I only needed two?
11-03-2010 02:17 PM
Which NBU version?
DBBACKUP used to show up in pre-6.x for Assigned media where a Media Server was not properly decommissioned (missing media server mediaDB entries).
11-04-2010 06:13 AM
We're running 5.0GA on solaris.
11-04-2010 06:26 AM
So - do you know if you have media servers that were decommissioned/reinstalled, etc that might have resulted in lost mediaDB entries? If you know the name(s) of the media server(s), you can use this TN to rebuild missing mediaDB entries:
Run the 'bpimagelist' with a start-date (-d mm/dd/yyyy) old enough that will ensure all image entries will be returned.
Any particular reason for running this Old, Unsupported NBU version?
11-04-2010 08:10 AM
I think we only have one media server and have not decommissioned or reinstalled it.
What got me started down this path was messages about backups failing because the robot had no available media. Our robot has room for 118 tapes and it was full. we export tapes for off site storage (for a year) and occasionally export tapes from diff backups for two months. So I was wondering why if the robot was full of tapes, why weren't any tapes available for the backups? I looked at the script to export the tapes, ran across the available media script, looked at the output from that and noticed quite a few tapes marked DBBACKUP and thought "that can't be right"
Is there anything I can do to make the oldest tapes available?
As far as the version of nbu, we've been bought and sold several times over the last few years and our current parent company isn't sure what to do about backups. :(
11-04-2010 09:03 AM
Not sure how familiar you are with NetBackup, so please excuse me if I'm stating something obvious to you.
Just because you have tapes physically in the robot does not necessarily mean they are available for use by NetBackup.
You must have tapes in "Active" (or "ActiveMultiplexed") Media Status in order for NetBackup to allow you to use them for backups. If they are in "Frozen", 'Suspended", or "Full" status, that means they will not be used for further backups until you manually change their status. In the NBU GUI, expand the "Media and Device Management" -> "Media" -> "Volume Groups", and check the "Media Status" column.
If you have tapes in "Active" status, check the Volume Pool name. If it is set to "NetBackup" (seems to be a default), then you may need to move them to the correct scratch Volume Pool. If the tapes are "Frozen", you need to determine why they are frozen (usually if your tape drives/robot are unavailable due to hardware failure or being offline and NBU can't mount a tape for a specified period of time, it freezes the tape; this can lead to all tapes in a library being frozen!).
If the status for the tapes says "Full", then you have tapes that still contain valid backup data on them. The NetBackup catalog keeps track of what data is located on each tape and when it is full, NBU doesn't scratch the tape untill ALL backup images have expired. You will need to determine if you just have tapes that should have gone offsite but didn't, or if you need more slots in your library to hold more data onsite.
11-04-2010 09:42 AM
Thanks Ron, I'm not very familiar with netbackup so your post is really helpful. If I can ask one more thing, what do you mean by "move them to the correct scratch Volume Pool"? What is the correct scratch volume pool?
11-04-2010 11:05 AM
If images are expired, the next step is to expire the media id in mediaDB on the media server before it gets deassigned in volDB. So, if the expiration cannot happen because of missing mediaDB entries, the deassign cannot happen - hence the false 'dbbackup' status in available_media output.
You will probably see losts of status 94's in Problem report or All Log Entries.
Please see this TechNote for instructions on how to fix this situation:
The expiration process is such that when there are no longer any valid images left associated with a specific media, the media is de-assigned from the mediaDB and the volDB so that it can become available again. The problem arises when there is a discrepancy in the volume pools between the mediaDB and the volDB to which the media belongs. Under these circumstances, NetBackup will retain the media in the databases and result in Error 94.
In short, you need to check if there are any valid images on the media-id's (bpimmedia), check for entries in all mediaDB's (bpimagelist), take appropriate action based on output found above, and them when you are 100% sure that there are no images associated with the media and no mediaDB entries, only THEN do vmquery -deassignbyid....
This will make tape available for backups again.
11-04-2010 12:11 PM
Thanks Marianne, I do not see and status 94's in the logs, but I do notice that bpmedialist will not print any info about the tape where vmquery will. I'll try the steps in the tech note and see if that helps.