There is a slight difference between tapes that are unlabeled (means no Networker label can be found - for example using brand new tapes or tapes overwritten with a different backup software) and tapes that are 'not in media index' (means they have a Networker label - this situation can be achived using nsrmm -d for example). Networker does not use tapes marked as 'not in media index' for automatic re-labeling operations when auto-media-management is enabled. Unlabeled tapes are used in this case.
I can understand ajw's problem. Esp. when you're cleaning up your Networker environment deleting tapes from networker's index may be nessesary to delete corresponding pool resources. But you'll have to relabel all tapes to a different pool so that these tapes can be re-used by auto-media-management-feature.
The only way is written to a tape with another program like tar, but you don´t need to remove the old volume to relabel tape to a new volume, Networker do this automatically.
Thanks I understand, was just trying to avoid having to manually label tapes in hopes that Networker would automatically make the tapes be unlabeled again. Thanks for your response.
Why would you remove tape in the first place if you wish to use it? One thing that comes to my mind is before where you had unique retention for clones and you wanted to keep clone longer, but in that case you would need to label manually such tapes which was logical thing. From ajw's post I can't see source and reason for the problem he run into and there might be another way of doing it what he does or what is happening to him.
I have deleted the volume from media db, but it doesn't work.
sometimes the error message is "label write, invalid argument" and sometimes is "RAP error: Tape label read for volume ? in pool ?, is not recognized by Networker: Input/output error"
I ve reinstalled all packages of the backup server, deleting the db before and it still doesn't work.
I ve tried to label the volume at another datazone and it works! (so the tape is ok)
NetWorker has a built in mechanism which will not relabel a volume which has a NetWorker label written to it if the NetWorker server does not know about this volume automatically. You can force a relabel of a volume from the command line using the appropriate flags including -Y. If it has been used by another software I am unaware of any situation where NetWorker will be unable to deal with this situation automatically. From your description so far I cannot tell what the issue is but would suggest the following: -
1. Can you mount another tape volume in a drive successfully?
2. If you can do this try to mount the volume you wish to relabel in this drive and see if you get the same error reported in your last post?
3. Try to relabel the volume from the command line using the -Y, -b -j flags (these are the minimum you need) specifying the pool you want and jukebox name.
I am not sure from your previous message if you are using the same volume when you say " you can relabel in a different datazone" as this would put a valid label onto the volume and you would hit my first point above.
There is a very good document available on Powerlink which details drivers etc for the different OS's we have, along with SCSI resets, CDI etc. This is called
Configuring Tape Devices for EMC NetWorker - search for this on Powerlink and you should be able to download, if you cannot find it put another post on and I'll see if I can find the exact link for you. It is maintained so it is best to check each time you need to reference this infornation as it may have beenupdated with later information.
As I ve tried to label the volume at another datazone (with another networker server and another tape library - same model) and it works! (so, the tape media is ok) the process was withdraw the tape from the strange datazone, deposit that tape at homologated datazone and right-click, "label" (bar code label), default pool and it works fine.
I ve tried already to use another tape drive inside that unstable library as you suggested, without success.
I ve tried to reboot the library, but doesn't work.
The command "nsrjb -L" sometimes labels temporarily and sometimes returns only error messages to all tape medias.
I'm starting to think the problem is at the sinergy of the hardware and scsi drivers.
Don't really know if this would apply to your situation, but we weren't able to label the tape with NetWorker when we have written to in the label segment of the tape with dd command (put the tape in device, dd'ed on tape without doing any rewinds/forwards).
To solve the issue, I took another NetWorker labeled tape in another tape drive, and used the same dd comand to copy the initial segment of the labeled tape over the dirty records on "damaged" tape... NetWorker then complained that it has two tapes with same label, however it would allow me to relabel the "damaged" tape with proper NetWorker label... perhaps this helps in your struggle .
dk3
163 Posts
0
April 29th, 2007 13:00
I can understand ajw's problem. Esp. when you're cleaning up your Networker environment deleting tapes from networker's index may be nessesary to delete corresponding pool resources. But you'll have to relabel all tapes to a different pool so that these tapes can be re-used by auto-media-management-feature.
axavier1
76 Posts
1
April 27th, 2007 11:00
ajw2
1 Rookie
•
88 Posts
0
April 27th, 2007 12:00
ble1
4 Operator
•
14.4K Posts
0
April 29th, 2007 06:00
ble1
4 Operator
•
14.4K Posts
0
April 30th, 2007 01:00
Monteiro_Victor
4 Posts
0
January 10th, 2010 04:00
I have the same problem. I'm trying to relabel old tape volumes (empty) used in the past by ARCServe.
It seems that ARCServe touch the volumes and put a flag that Networker doesn't understand.
Networker can't read the old label and data (expected) but can't relabel it.
There is any way to force a deletion/format this tape volumes?
Monteiro_Victor
4 Posts
0
January 10th, 2010 23:00
I have deleted the volume from media db, but it doesn't work.
sometimes the error message is "label write, invalid argument" and sometimes is "RAP error: Tape label read for volume ? in pool ?, is not recognized by Networker: Input/output error"
I ve reinstalled all packages of the backup server, deleting the db before and it still doesn't work.
I ve tried to label the volume at another datazone and it works! (so the tape is ok)
The command "nsrjb -L" doesnt work!
ark-flUcr
23 Posts
0
January 10th, 2010 23:00
masonb
445 Posts
0
January 11th, 2010 00:00
NetWorker has a built in mechanism which will not relabel a volume which has a NetWorker label written to it if the NetWorker server does not know about this volume automatically. You can force a relabel of a volume from the command line using the appropriate flags including -Y. If it has been used by another software I am unaware of any situation where NetWorker will be unable to deal with this situation automatically. From your description so far I cannot tell what the issue is but would suggest the following: -
1. Can you mount another tape volume in a drive successfully?
2. If you can do this try to mount the volume you wish to relabel in this drive and see if you get the same error reported in your last post?
3. Try to relabel the volume from the command line using the -Y, -b -j flags (these are the minimum you need) specifying the pool you want and jukebox name.
I am not sure from your previous message if you are using the same volume when you say " you can relabel in a different datazone" as this would put a valid label onto the volume and you would hit my first point above.
Regards,
Bill
masonb
445 Posts
0
January 11th, 2010 09:00
Configuring Tape Devices for EMC NetWorker - search for this on Powerlink and you should be able to download, if you cannot find it put another post on and I'll see if I can find the exact link for you. It is maintained so it is best to check each time you need to reference this infornation as it may have beenupdated with later information.
Regards,
Bill
Monteiro_Victor
4 Posts
0
January 11th, 2010 09:00
As I ve tried to label the volume at another datazone (with another networker server and another tape library - same model) and it works! (so, the tape media is ok) the process was withdraw the tape from the strange datazone, deposit that tape at homologated datazone and right-click, "label" (bar code label), default pool and it works fine.
I ve tried already to use another tape drive inside that unstable library as you suggested, without success.
I ve tried to reboot the library, but doesn't work.
The command "nsrjb -L" sometimes labels temporarily and sometimes returns only error messages to all tape medias.
I'm starting to think the problem is at the sinergy of the hardware and scsi drivers.
mofo
58 Posts
0
January 12th, 2010 07:00
Don't really know if this would apply to your situation, but we weren't able to label the tape with NetWorker when we have written to in the label segment of the tape with dd command (put the tape in device, dd'ed on tape without doing any rewinds/forwards).
To solve the issue, I took another NetWorker labeled tape in another tape drive, and used the same dd comand to copy the initial segment of the labeled tape over the dirty records on "damaged" tape... NetWorker then complained that it has two tapes with same label, however it would allow me to relabel the "damaged" tape with proper NetWorker label... perhaps this helps in your struggle
.
cheers.
Monteiro_Victor
4 Posts
0
January 12th, 2010 11:00
Thanks for your help!
I have not used dd commands, I'll run a grab to send to emc support.
I will run "inquire" and "sjisn" commands too.