05-17-2012 06:11 AM
Hello everyone,
I've just strated using Netbackup in an effort to recover data stored on some LTO2 tapes. Unfortunately I'm running into the following problem whenever I try to inventory the tapes in order to be able to perform the 2 phase import process:
Unknown or corrupt media detected in drive
I'm currently using NetBackup 6.5 on Windows 2008 Server (32bit) and a standalone tape drive.
Is there any way to get vmphyinv to inventory the tape even though I'm getting this error?
Thank you for your help and time,
-Jorick
05-21-2012 08:28 AM
If NBU can read the header it "should" then adjust to the block size that the tape was written with.
If you are running on Windows 2003 / 2008 then it should be able to use a 256kb block size - but if anything in the HBA driver does not permit this then it could cause the issue potentially.
As you are on 2008 you should be OK - but worth checking what the drivers registry entry says for the HBA you are using.
Identify the driver used - go to HKLM\System\CurrentControlSet\Drivers\driverused\Parameters\Device
and look for the MaximumSGList key
I believe it needs to be 65 decimal or 41 hex - any lower and it will not work - high should be OK (will need to support it to be changed as well as needing a reboot)
Hope this helps
05-21-2012 09:45 AM
05-21-2012 09:45 AM
05-22-2012 05:34 AM
Hi everyone,
I've changed the block size and I'm not getting any results. I've tried randomly selecting tapes among those that I have and i've found that they all start like this:
05-22-2012 05:48 AM
sorry let me ammend. Only up to 300 is it the same on all tapes
05-22-2012 06:21 AM
Hi Jorick ...
Even a spanned tape will have a normal netbackup media header.
So if I understand correctly :
(1)
We cannot see a proper NBU header at the beginning of the tape (this WILL stop NBU reading the tape )
(2)
As you skip through the tape, you can read the NetBackup Backup header
...
very very odd .. it's looking like the header was incorrectly written and you only find out later ( like now) , but, this would have meant that every tape you found MUST have been written in one session (else it wouldn't have been able to remount them).
We have certainly seen some odd things with tape drive firmware, I wonder if this is another one ...
Marianne/ Mark - do you agree with me ?
Thanks,
Martin
05-22-2012 06:31 AM
Marianne/ Mark - do you agree with me ?
Yes, I remember a very lengthy thread not-so-long-ago that was solved with updated firmware.
I actually stopped respondig to this thread after noticing that NBU media server was installed on 32-bit W2008.
Customer tried the same at my previous company. Device management service would not start after device config. Tried to troubleshoot for days.... Found the 32-bit compatibility issue.
Customer re-installed OS - this time 64-bit (was already 64-bit server). Installed 64-bit NBU and everything worked.
I am surprised to see all working fine with LTO4 media.
The other thing that comes to mind is if the tape is REALLY LTO2.
Any possibility it could be LTO1 media that was used in LTO2 drive?
Tape may even have LTO2 label on LTO1 tape....
05-22-2012 06:42 AM
If the media was incomatable with the drive, it wouold never load ...
05-22-2012 06:57 AM
Also true for StandAlone?
I have once seen a customer who tried to write to LTO1 tapes in LTO3 drives. They thought the 'clever' thing would be to put LTO3 labels on.
Robot loaded the tape and only reported an error when trying to read header.
05-22-2012 07:15 AM
Oohhh, I stand corrected .. intersting ...
I always though that the tape type was on the chip in ythe cartridge, and therefore it was inpossible to load the wrong tape into he wrong drive ...
Clearly not ...
M
05-22-2012 07:17 AM
that in the example M has quoted you *would* be able to load an LTO1 into an LTO3 drive....
...whereas initially, a few posts back, looking at the possibility that an LTO1 tape with an LTO2 label was being loaded into an LTO4 drive....phew...
05-22-2012 07:40 AM
Wow this is getting interesting. So here are some more details as to what is going on now. I am sure that the tapes are LTO2 as they are from HP and HP LTO2 tapes are dark red each version of LTO in HP has a different color.
I wasn't around for the initial extract of the LTO4 tapes but I did assist in the later extractions and as I said they worked fine. I do however know that they apparently had to change to W32-bit and *downgrade* the firmware on the drive that we were using to read the LTO4 tapes.
If I can find a copy of W2008 64-bit I will install that and give it a shot in the mean time here are the file headers I've been reading from the tape using nt_ttu fsf to position myself and then read the first 1024 bytes:
Translator at: http://home.paulschou.net/tools/xlate/
Tape:
File 1:
File 2: