I have a question about nearline volumes.
We eject all our full tapes from library and also remove them from CAP. So the tapes are not in there physicaly. We have a script which is updating locations, but first it finds out whether or not we have tape in a library (via -q near). If the tape is in library, but the location differs, it prints out an error message.
The problem is we are still getting these error messages, because when i run
"mminfo -q near -xc, -r volume " I still have these volumes on the list even if they are not physicaly in the library... Our offsite storage guys confirmed, that tapes arrived and are there. So its no way we can have them in library.
Any ideas why is this? How is it possible? And how can I convince our library to show only the tapes that are really there? how mminfo determines whether the tape is in library or not (maybe we have some problem somewhere else)?
we use the same script on 6 libraries, on 5 everything is ok, but this one is weird.
networker version is 7.4.2.Build.431
we have volume priority set as NearLine Priority
This can happen even though the tape volumes are taken out from the library, the location field on the tape volume might not be updated , You can verify by running the command
mminfo -q "volume=XXXX" -r location - Replace XXXX to a volume name where you are facing issues.
If still it shows incorrect , Then You may have to do media scavenging to get rid of this issue.
yes, I can see the location was not updated, nor the nearline flag.
Question is, why is it happening. Because its not one tape, we have about 1500 volumes nearline (but only about 700 physicaly in library), seems to me we have this problem for a while.
With scavenging, we need to wait until next scheduled reboot, we cannot stop services on our backupserver without good cause.
But thank you for your answer
During scavenging , its not required to reboot the server , Its just we have to restart nw services once or twice.
Once you take out the volume from the Library , the nw should update the current location , which is not happening in your case.
It could be because of media database corruption.
I know that I dont need to reboot server during scavenging. But we are not permitted to restart nw services unles in scheduled maintanence window, which is only when server reboot is needed.
You can make an update to location with the command
mmlocate -un volume location
You can put in a script to automatically remove the tapes from the library and then update the location in the media database.
volume = x
nsrjb -w $x
mmlocate -un $x location
I know how to update location via mmlocate. We have a bit complex script for that, but doing the same. Problem is, that we need to check if the tape is already in library and this is done via mminfo -q near -r volume. And we want to update only the tapes, that are not in library. But unfotunately our library is not removing the near flag from media db when doing
nsjrb -w tape
which should be done automaticaly
and so the script is not working...this is our problem, not updating the locations, but the near flag in media database.
Based on a Networker-online forum discussion from a while back, I think that what you are looking for may not be
currently possible with NetWorker. There is an 'offsite' flag that can be set for a volume, but there is no way to query
You can declare a volume 'offsite' with:
nsrmm -o offsite <volume>
This flag is then used internally in order that NetWorker can know that it is offsite and take this
into consideration for volume selection for cloning, recover and staging.
However, you cannot get a list of 'offsite' volumes from mminfo.
An RFE (reference LGTsc07449) has been filed for this and it is being considered by the Product Development team
so that users could be able to identify all 'offsite' volumes easily. Maybe the 'offsite' flag can help you anyway
even without being able to query it. If you are interested in having this functionality added, you should open a
Service Request and ask to have your details added to the RFE LGTsc07449.
I agree - problem is mminfo is used to determine if volume is nearline or not. There is number of possible situation when media database might not be updated to correctly reflect location status. So in your case, I would rather use sjirdtag command to determine if volume is withing library or not. Doing sjirdtag should give you the list of all volumes within library - all those volumes in mdb which are not listed would not be nearline.