Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations strongm on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Networker 7.0 doesn't unmount tapes during Inventory

Status
Not open for further replies.

rleflore

IS-IT--Management
May 30, 2003
20
US
I've recently upgraded to version 7.1 of Networker. Since upgrading I've noticed that during an Inventory the software fails to unmount mounted volumes.
i.e.
If I inventory slots 1-83 and there is a tape in one of the drives that is loaded in slot 3; I then execute the Inventory of all slots from using any drive, it will fail at drive 3 with a GUI message stating "Slot 3 in in drive #", Inventory finished. (This is true for all drives in the machine)

I have to unmount slot 3 first in order to do the inventory for slots 1-83 or do the inventory in 2 parts. Inventory 1-2 and then 4-83. This was not a problem in the previous version. Is there a fix for this or is this just the nature of the newest version?
 
Can't say for 7.1 but for 6.1.2 an inventory of the slots don't umount tape/s still in the tape-drive - nsrjb -HE would unload tape in the drive.
 
From which version did you update? Legato has changed the tape driver from 6.x to 7.x. The new driver type is called Common Device Interface (CDI) and it is enabled by default. However, there are some restrictions. Please consult the release notes for details.

You may want to try the old control mechanism. To achieve that, open the device resource and change the Arribute "CDI" from "SCSI commands" to "not used".

 
i've reset all drives to "not used" and tested the opertions, however they still function incorrectly. To answer your question, the previous version was 6.0.1. Also, I've checked the release notes and there is no reference to this being a loss of functionality.

To reiterate, the old version wouldn't unmount the tapes but would, and would at minimal at least acknowledge tapes in the drives and proceed to the next slot.
I'm using sol8, Sun450, 1.5 gb mem,. QUANTUM ATLM2500, 83 slots, fibre channel, 4 SDLT 320 Quantum drives. Using new fuji SDLT medium, no recycles. If this is the nature of it, then I can get by with a script that will do it for me.
 
I do not think that this is a general NetWorker problem - if this would be the case, Legato would have fixed it pretty soon. I personally would think that the eject option (not unmount) is controlled by the jukebox and not by NW (usually you can set this up in the jukebox parameters) - i
just will not understand why this should happen when you only update the software. Anyway, it would be worth to investigate.

Also you could run jbexercise to test the jukebox. The description is in the manpages.
 
With 7.1.1 on Solaris, if I do an nsrjb -I it unloaded the
tapes on the library.. This is an ATL P3000 direct attached.
 
By "unmount", do you mean the status of a networker volume as "mounted" or "unmounted" logically (seen via nsrwatch), or are you talking about physical unmounting performed by your silo?

For logical unmounting, my 6.1.4 install will unmount/eject any drive that isn't busy if I tell it to do a full inventory. Physical failures to eject are usually in my case (an IBM 3494 populated with 3590s) hardware issues.
 
Just for clarification:

NW must recognize a NetWorker label in its own media database to be able to mount the media. In other words:

- The media must be labeled by NetWorker
- It must exist in the database (it must have been either
labeled or scanned by this NW server)
If an unknown NW media has been detected, it can not be mounted.

I think for labeling, we should use load/unload instead of mount/unmount.
 
Referring back to the original post:

"If I inventory slots 1-83 and there is a tape in one of the drives that is loaded in slot 3; I then execute the Inventory of all slots from using any drive, it will fail at drive 3 with a GUI message stating "Slot 3 in in drive #", Inventory finished. (This is true for all drives in the machine)"


This is a known issue with 7.0 and 7.1. Problem did not happen in 6.x.

This problem is scheduled to be fixed in NetWorker 7.1.2 later in 2004.
 
Good to know this for when I consider going to 7. Right now I just do this anyway:

nsrjb | grep -- "\-\*" | awk '{print $3}' | while read T
do
nsrjb -I -T ${T}
done

In my case I usually don't want networker using all my devices for inventorying anyway, 1 is fine so I can use my other drives to get other work done.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top