recover command from client not automatically mounting media

Could it be that the jukebox has been set to "service mode" or it has been "disabled" ?

Hi guys,

  firstly, the library is working hence I can manually mount the tapes via nrsjb. The library status is:

Jukebox SL500: (Ready to accept commands)

Regarding read hostname etc, the environment is a single Networker server and a single Windows Server client (network client not a storage node). So I don't there should be any problems here.

The tapes themselves were long expired so I'm rebuilding the indexes via "scanner -i" so again I don't think there should be any issue with pools being linked to devices although I'm not sure where I would check that. Unfortuately there is no GUI installed on the server so I'm stuck with command line, nsradmin etc. I've managed to do what I needed via nsradmin to this point but its pretty hard going if you haven't any previous experience of it, which I hadn't!

thanks, Andy.

Wow – Andy 7.3 – a blast from the past!

From memory I think it should also work although I do recall there was some circumstances that a manual mount would be required.

Is you backup pool linked to certain devices?

Can you check the read hostname value of the jukebox and see if this is set and if so is it set to a storage node which you need or is allowed to mount the pool?

If it is not set at all – set to the storage node you want to use and then retry the restore.

Regards,

Bill Mason

Andy – nsrwatch should be available to you to see the devices etc configured.

As for nsradmin you can see the jukebox with the following:

Show all

Print type: nsr jukebox

You should be able to see what the read hostname value is (it used to take the name of the storage node with owned the first device) which in your case as there is only the NetWorker server with devices should be set to that (please check the exact name of the variable).

If not type a “.” After the query above and then

Update Read hostname:

Then answer y to prompt.

Regards,

Bill Mason

Thanks Bill.

  I do indeed have nsrwatch to hand.

I've checked the value of read hostname, it was null previously. I've just changed it to the Networker server name but this doesn't seem to have made any difference. I've kicked off another restore from the same Windows client and I just see this in nsrwatch:

server2: browsing

Messages:

Thu 15:57:35 /dev/rmt/3cbn mounted LTO Ultrium-3 tape tape.022 (write pro

Thu 15:57:35 ctc:L:\Scor (3/19/09) starting read from tape.022 of 792 G

Thu 19:29:07 server2: browsing

Thu 19:29:08 server2: done browsing

Thu 19:29:08 server2: browsing

Thu 19:29:10 media waiting event: waiting for LTO Ultrium-3 tape tapeBD.001

Pending:

Thu 19:29:09 media waiting: waiting for LTO Ultrium-3 tape tapeBD.001 on

So same old story,

thanks, Andy.

Bill's statement: "Is you backup pool linked to certain devices?" could point into the right direction.

If you start the scanner command, NW will tell you about the volume, the pool and whether the pool exists.

Pipe the output to a file for verification.

Now, if this pool does not exist, create it.

Since a while, you can even do that without any selection criteria - if not, they would not matter as you do not use it for backups.

Next, enable your pool for the use of all tape devices. Now NW should be able to load and mount the tape in any drive.

And this makes the difference:

  - recover needs a loaded and mounted tape.

  - scanner 'only' needs a loaded tape. This is important as otherwise NW would not be able to read an unknown tape.

If you have not enabled your devices for all pools, you will be able

  - to load the media

  - to label the media

but not to mount the media later.

Andy__Smith wrote: Hi Hrvoje,   in what respect does it seem I do not use a barcode scanner? And yes the library has been inventoried and is correct and up to date. I've checked both tape devices and they are set as type "LTO Ultrium-3".

tape.022 is not the volume name you get when you use barcode to match volume names.  You can do something like nsrjb -V to get scsi address on which library is and then you can do sjirdtag b.t.l (but.target.lun) to see barcodes.

Hi Bingo,

  sorry, maybe I've not been clear with the terminology. The issue is that the jukebox does not automatically move the tape to the drive. When I mounted the tape manually I did a nsrmm -m -f /dev/rmt/3cbn to "mount" the tape, this is expected and not an issue.

What I want and expect is if I run a recover on the Windows client that it should automatically move the tape within the jukebox from slot X to a free drive without any further intervention. This would be normal would it not?

thanks, Andy.

Hi,

  ok, I´ve checked the pool. The pool for these tapes does exist, is enabled, is of type "Backup" and has devices set to:

devices:  /dev/rmt/3cbn    /dev/rmt/4cbn ;

which is correct,

thanks, Andy.

Hi Hrvoje,

  in what respect does it seem I do not use a barcode scanner? And yes the library has been inventoried and is correct and up to date. I've checked both tape devices and they are set as type "LTO Ultrium-3".

Hi Bingo,

   I am not next to the library either so cannot label a tape as the it isn't loaded with any spares currently. With respect to jbexercise, is this necessary given I know I can load/unload tabes via nsrjb? That is mechanically its fine and I can use Networker tools to operate it,

thanks, Andy.

Hrvoje Crvelin wrote: tape.022 is not the volume name you get when you use barcode to match volume names.  You can do something like nsrjb -V to get scsi address on which library is and then you can do sjirdtag b.t.l (but.target.lun) to see barcodes.

Ah ok, yes the barcode does not match the label. This is how the tapes were created, I'm just trying to recover the data off them. The jukebox inventory has to mount every single tape due to this issue, which it has done.

thanks Andy.

Well - one step after the other.

My obvious assumption was that the restore is more important.

Unfortunately, i cannot really help you on the jukebox issue as i am not sitting next to it. But next i would try to ...

  - label another tape

  - use jbexercise to verify the jukebox operation.

I noticed that you don't use barcode reader and I remember 7.3 had issues with that.  With that in respect, can you confirm that library is inventoried and that nsrjb database is up to date?  Also, are current devices also configured media type LTO3?

o.k. - i think i found it.

In your opening message you state:

  ....

02/27/14 15:56:30 nsrd: [Jukebox `SL500', operation 1]. Initiated operation `nsr jb -ln -S 47 -f /dev/rmt/3cbn'.

The last line being where I manually ran a mount with nsrjb.

No.     '-ln'  simply loads the tape without mount. Use  '-l'  instead.

I think what you might have is bug from 7.3 - 7.3 had so many changes that it took couple of services packs to make it stable.  From what it seems, one that you have is when subsequent tape has to be loaded that doesn't happen. I'm surprised no one did update of that version. At this point, you are left with workaround which is to mount it manually.