Unsolved
This post is more than 5 years old
4 Posts
0
12270
TL2000 write failure
TL2000, Poweredge 6850, Redhat Server 5.2 (vmlinuz-2.6.18-53.1.4.el5)
After having trouble getting the standard tapeware softare (yosemite) working
with my tape drive (just one single tape in the drive) using a standard dell server with a standard dell scsi card with
a standard Redhat O/S (dell infact selected all the components for me), I tried to just get it up and running using the mt and tar commands.
I've upgraded the tape drive, scsi card and O/S software (including the driver) already to avoid software bug
issues. I think the problem is most likely with the operator ? me !
Anyway, after testing the tape drive with IBMs test software, I learned that this drive
will only operate with tar -b 64 or higher. I've been using tar -b 256 (or 512).
After setting
mt -f /dev/st0 setblk 512
mt -f /dev/st0 compression 0
(to avoid issues associated with block size and compression)
I used the basic application of
tar -f /dev/st0 -b 512 -c ./foldername and
tar -f /dev/st0 -b 512 -vt or tar -f /dev/st0 -b 512 -x
and everything seems to work fine.
HOWEVER.
If I try and backup a large amount of data, the backup runs successfully
for a few hours (not nearly enough data to fill up the tape) and then crashes every time
with this error in the /var/log/messages file.
Dec 2 13:12:37 mrsc1 kernel: st0: Current: sense key: Aborted Command
Dec 2 13:12:37 mrsc1 kernel: Add. Sense: Nak received
Dec 2 13:12:37 mrsc1 kernel:
Dec 2 13:12:37 mrsc1 kernel: Info fld=0xc9, ILI
Dec 2 13:12:41 mrsc1 kernel: st0: Current: sense key: Unit Attention
Dec 2 13:12:41 mrsc1 kernel: Add. Sense: Power on, reset, or bus device reset occurred
Dec 2 13:12:41 mrsc1 kernel:
Dec 2 13:12:41 mrsc1 kernel: st0: Error on write filemark.
Many of you know much more than I do in this area, and I'd appreciate any insight
as to what is causing this failure?????
Thanks in advance,
Dave Powell
DELL-Bob D
899 Posts
0
December 5th, 2008 15:00
does it crash roughly at the same time everytime?
Have you tried setblk 0 instead of a fixed block?
You mention that not nearly enough data to fill the tape.. I'm assuming LTO.. what generation? Have you tried the -M on the tar to create multi-volume backup just incase it is running out of tape for some reason?
...seems I've heard of this before but just can't recall the solution.. I'll dig through some old posts to see if I can dig up anything.
What was up w/ Yosemite that you moved onto tar?
Bob
dkpowe00
4 Posts
0
December 16th, 2008 11:00
Hi Bob,
I appreciate you responding. I still have not been able to get this tape drive
to function properly. It did crash roughly at the same place every time. I am able
now to get it to write to the tape drive basically if I use the -b option with tar and force
the block factor to be 64, 128, 256 or 512. The setblk was set to 0 by default.
I fiddled around with setblk and compresion, but basically it seems I need to use the large blocking
factor in tar to get it to write.
I can not however read the tape. Basically, I get some sort of error no matter what I do
when trying to read the tape. I can use the exact same tar command (with x instead of c)
using the same blocking factor and can not read the tape. I get an input/output error to
the screen and a current: sense key: aborted, add. sense: NaK received to /var/log.
I download and installed IBM's Tape diagnostic too (itdt). It can see the drive. When I
run Test is can sucessfully write at 512kb (100 MB/s), 256kb(100 MB/s), 128kb(100 MB/s),
64kb(80 MB/s), and 32kb(5 MB/s). It aborts the rest of the test and can not write at 16kb
blocks and lower (This is consistent with me not being able to use tar unless I set the
blocking size to something other than its default (20)). When I run the dump :emotion-46: test on
IBM's software it FAILS! The generated log says "Read Buffer Failed at offset 0".
Is my drive bad? any help would be appreciated. I'll try anything you suggest.
we haven't got this drive to work since we bought it.