Hi, It could be a media problem-I have faced this kind of issues in SDLT Tapes and I have had contacted media vendoe FUJI - and found out that moost of these tapes are gone bad due to improper handling.this include -Sending tapes for offsite without proper packaging -Tapes kept in dusty area -Tapes kept vertically Also check cleaning catridege should not use more than 20.
Hi, Lot of our LTO-3 Tapes are marked "full", using 10 or 20 Gb or less (other tapes have up to 400 Gb) and the daemon.log shows IO errors writing on theses tapes, on both drives. We changed drives, scsi cables, scsi terminator without success. Is this a hardware problem? Is it a good idea to try to reduce Scsi speed? and how? (Networker 7.1.4/Windows2K)
This could be a tape drive problem only. It would be difficult to assume all media cartridges to go bad simultaneously.
One good option could be cleaning the drives but if you say you have allready replaced the drives, try and get a couple of fresh tapes.
Also, if possible - upgrade to NW 7.2.2. Are you using the correct media type and did the backups ever work good in this setup?
The backups worked fine last month. Now I will try a couple of cleanings.
When I look in library status (MSL6030) , I can see these message: Drive error: No Media error: Yes Daemon.log says: Drive status:drive status is Not Ready, ASC 0x3a ASCQ 0x04 What does it means?
This is non standard ASC/ASCQ code so real meaning should be searched somewhere in librarary docs (usually they have table of generally not implemented ACS(Q) codes that library uses). What library seems to suggest to is that there has been some kind of error with those medias - knowing what exactly that ACS(Q) means could help so I would suggest to ask that library vendor.
I have been having a similar problem with my LTO-3 tapes with Networker 7.2.1. My 2 drives are in a Dell ML6000 library connected via an EMC Clariion SAN to a Win2003 server. I have the drives persistently bound too through the fibre.
Everything is fine until I have to reboot the server. Then kaboom! I start getting lots of errors with media verification and "Read-open error:no tape in drive" which causes tapes to be marked Full at like 20 GB. I have used new media, no difference, totally random. Media verification even fails halfway through a job! I'm about to lose my mind. Help please.
Your problems seems more as something else (so prehaps separate thread would be good idea). Judging by error it looks as OS paths have changed so you may wish to recheck that.
NetWorker cares about 2 thing: - OS path which is something like /dev/rmt/3mnb or whatever - physical device order which can be obtained via sji tools
When you run jbconfig order of devices must be configured according to physical order. So, if 1st drive is 3mnb, 2nd one 1mnb and 3rd one 2mnb that doesn't give you sorted order, but that is the right way of doing it. Your current physical order defined in NW is the one you see in nsrjb output at the end. To verify it with current physical order in library use mt command for each drive.
The tapes are in the correct order. I just reinstalled the jukebox and tapes into Networker and right away when doing inventory it got a "Unload retry for 'ML6000' failed" error. Uhhh...its frustrating. Its like the Jukebox and Legato are out of sync or something.
Make sure you do NOT have any tape loaded. And the follow following steps (I will assume you have tape in first slot): 1. sjirdtag t.b.l (scsi address for jukebox) 2. sjimm t.b.l slot 1 drive 1 3. mt -f /dev/rmt/foo status (where foo is what you bealieve is the first device, but you better do it for all - only the one with tape loaded will give you different output indicating tape is loaded) 4. mt -f /dev/rmt/foo offline (where is foo the device that was identified to have tape) 5. sjimm drive 1 slot 1 6. sjimm slot 1 drive 2 7. repeat step 3 8. mt -f /dev/rmt/foo offline 9 sjimm drive 2 slot 1 10. etc... (do the rest until you get to the last drive)
Above will give you physical drive order. Then, jbconfig should be done according drive order above (1,2,3, etc) and answers in jbconfig should be according to OS paths identified.
Something I have noticed is that the drive order changes sometimes In Windows if I reboot or uninstall the drives from Device Manager. It reverses them. And then when I do mt -f status it shows the drive 1 now being what was formerly drive 2. I guess its more windows than anything. Thanks for the help.
saleesh1
128 Posts
0
November 20th, 2006 09:00
It could be a media problem-I have faced this kind of issues in SDLT Tapes and I have had contacted media vendoe FUJI - and found out that moost of these tapes are gone bad due to improper handling.this include
-Sending tapes for offsite without proper packaging
-Tapes kept in dusty area
-Tapes kept vertically
Also check cleaning catridege should not use more than 20.
Thanks
Saleesh
amediratta
2 Intern
•
2K Posts
1
November 20th, 2006 08:00
Lot of our LTO-3 Tapes are marked "full", using 10
or 20 Gb or less
(other tapes have up to 400 Gb)
and the daemon.log shows IO errors writing on theses
tapes, on both drives.
We changed drives, scsi cables, scsi terminator
without success.
Is this a hardware problem?
Is it a good idea to try to reduce Scsi speed? and
how?
(Networker 7.1.4/Windows2K)
This could be a tape drive problem only. It would be difficult to assume all media cartridges to go bad simultaneously.
One good option could be cleaning the drives but if you say you have allready replaced the drives, try and get a couple of fresh tapes.
Also, if possible - upgrade to NW 7.2.2. Are you using the correct media type and did the backups ever work good in this setup?
amediratta
2 Intern
•
2K Posts
1
November 20th, 2006 08:00
This also suggests media problem. Go ahead for a new fresh seal packed tape.
The ASC error you mentioned should be temporary, may be when you opended the library physically Legato might have sensed an error condition.
jlcarton
1 Rookie
•
10 Posts
0
November 20th, 2006 08:00
Now I will try a couple of cleanings.
When I look in library status (MSL6030) , I can see these message:
Drive error: No
Media error: Yes
Daemon.log says: Drive status:drive status is Not Ready, ASC 0x3a ASCQ 0x04
What does it means?
ble1
4 Operator
•
14.4K Posts
0
November 20th, 2006 12:00
DeaconZ28-2015
252 Posts
0
November 24th, 2006 06:00
Everything is fine until I have to reboot the server. Then kaboom! I start getting lots of errors with media verification and "Read-open error:no tape in drive" which causes tapes to be marked Full at like 20 GB. I have used new media, no difference, totally random. Media verification even fails halfway through a job! I'm about to lose my mind. Help please.
ble1
4 Operator
•
14.4K Posts
0
November 24th, 2006 07:00
DeaconZ28-2015
252 Posts
0
November 24th, 2006 07:00
What are OS paths? Should I bind the drives by Port ID instead of World Wide Port Name?
ble1
4 Operator
•
14.4K Posts
0
November 24th, 2006 07:00
- OS path which is something like /dev/rmt/3mnb or whatever
- physical device order which can be obtained via sji tools
When you run jbconfig order of devices must be configured according to physical order. So, if 1st drive is 3mnb, 2nd one 1mnb and 3rd one 2mnb that doesn't give you sorted order, but that is the right way of doing it. Your current physical order defined in NW is the one you see in nsrjb output at the end. To verify it with current physical order in library use mt command for each drive.
DeaconZ28-2015
252 Posts
0
November 24th, 2006 09:00
ble1
4 Operator
•
14.4K Posts
0
November 24th, 2006 10:00
Make sure you do NOT have any tape loaded. And the follow following steps (I will assume you have tape in first slot):
1. sjirdtag t.b.l (scsi address for jukebox)
2. sjimm t.b.l slot 1 drive 1
3. mt -f /dev/rmt/foo status (where foo is what you bealieve is the first device, but you better do it for all - only the one with tape loaded will give you different output indicating tape is loaded)
4. mt -f /dev/rmt/foo offline (where is foo the device that was identified to have tape)
5. sjimm drive 1 slot 1
6. sjimm slot 1 drive 2
7. repeat step 3
8. mt -f /dev/rmt/foo offline
9 sjimm drive 2 slot 1
10. etc... (do the rest until you get to the last drive)
Above will give you physical drive order. Then, jbconfig should be done according drive order above (1,2,3, etc) and answers in jbconfig should be according to OS paths identified.
I would start from that.
DeaconZ28-2015
252 Posts
0
November 27th, 2006 09:00