Dat drive problem, wont read

[David Preston_4]

Hi I have a k200 wit a dds2 dat drive. I can write to the tape, but I can not read without a directory checksum error. I have replaced the drive, the Core I/O controler, and even the scsi cable. and it will not read (tar -t) I have also used fbackup / frestore with no success. Does anyone have any suggestions???

Dave

[Uday_S_Ankolekar]

Hi,

Did you rewind the tape before read..?
If not try with mt command
mt -f /dev/rmt/omn rew

than do a tar -tv

Goodluck,
-USA...

[Marco Paganini]

Hello

Use the /dev/rmt/0m device. It should rewind the tape automatically for you. Also, try to *clean* your DAT drive with a cleaning cartidge! The drive heads need to be cleaned or else you'll have all kinds of problems.

Regards,
Paga

[Darrell Allen]

Hi David,

Has this just started happening? Perhaps you can try specifying the tape drive:
tar cvf /dev/rmt/0m /tmp
tar tvf /dev/rmt/0m

Are you sure the drive's device path hasn't changed perhaps due to hardware added?

Darrell

[A. Clay Stephenson]

Hi David:

Although you did not mention I assume that you have tried different tape. What I want to do is get tar, fbackup, etc. out of the picture and talk to the drive directly. Try this:

dd if=/stand/vmunix of=/dev/rmt/0m (or whatever). This may take some time as you are writing in 512-byte blocks.


Next dd if=/dev/rmt/0m of=/tmp/xx

Now let's compare the two files.

cksum /stand/vmunix
cksum /tmp/xx

If they differ, I want you to use od to find where things go wrong. It will proably be on a block boundary.

The only other thing that I can think of is bad termination. I can't remember on these boxes if the terminator is built on to the end of the internal cable or whether you have to enable terminal on the last device on the bus. Also, is terminator power enabled on at least one device on the bus? A bad terminator would do exactly the things that you describe. The other thing that would cause this is termination being applied to anywhere except the ends of the cable.

Happy hunting, Clay

[David Preston_4]

OK, I have to admit that I did not give you all the facts, so here it is from the top,
I picked up the machine used, and it had previously had this issue, I have never gotten a tape to properlyu read or recover from this device. 1st I replaced the scsi cable (including terminator) no change. 2nd I changed out the drive with a new one, both drives were able to write a tape I could read on another system. 3rd I changed out the Core IO, all parts have come from a reliable vendor so I would think that the hardware is fine, the only pieces of hardware not checked are the interface panel on the back of the disk cabinet, but that is basically a pass through, and the backplane, which frankly if thats what is wrong I am out of luck. the only consistant thing is that the error always says "directory checksum error" but it happens in a different place each time. every copy I have done has been in a directory of multipl files, I have not tried a single file.
Other than that I don't know what to tell you.
Is it still worth doing that block write? what can I deduce from the stoppiong place of the recovery?

Dave

[A. Clay Stephenson · Accepted solution]

Okay Dave:

This is now sounding much more like a software problem than a hardware problem. Have you checked for all the latest SCSI/tape patches?
I really suspect a problem with the stape driver.

[Sanjay_6]

Hi David,

Can you post your ioscan output. Also Do you have proper driver configured in the kernel. What is the syntax you are using to copy data and read data from the tape.

Regds

[David Preston_4]

Hi, thanks for all the feed back:
here is the ioscan:
ioscan
H/W Path Class Description
===============================================
bc
8 bc I/O Adapter
10 bc I/O Adapter
10/0 ext_bus GSC built-in Fast/Wide SCSI Interface
10/0.0 target
10/0.0.0 disk HP C2490WD
10/0.3 target
10/0.3.0 disk HP C2490WD
10/0.4 target
10/0.4.0 disk HP C2490WD
10/0.5 target
10/0.5.0 disk HP C2490WD
10/0.6 target
10/0.6.0 disk HP C2490WD
10/0.7 target
10/0.7.0 ctl Initiator
10/4 bc Bus Converter
10/4/0 tty MUX
10/4/8 tty HP J2094A - 16 Modem MUX
10/12 ba Core I/O Adapter
10/12/0 ext_bus Built-in Parallel Interface
10/12/5 ext_bus Built-in SCSI
10/12/5.0 target
10/12/5.0.0 tape HP C1533A
10/12/5.2 target
10/12/5.2.0 disk HP DVD-ROM 6x/32x
10/12/5.7 target
10/12/5.7.0 ctl Initiator
10/12/6 lan Built-in LAN
10/12/7 ps2 Built-in Keyboard/Mouse
32 processor Processor
34 processor Processor
49 memory Memory
#
also attaching text file

how do I find out about the driver software?

Dave

[A. Clay Stephenson]

Probably the best answer at this point is to install the latest General Release patchset.

[Frank Slootweg]

Have you checked the syslog.log and dmesg(1M) output for errors? Note that both are buffered, so you may see the errors only some time after they occured.

Also please post a failing tar(1) example (assuming that the "directory checksum error" is from tar). Please post the full commands, options, parameters used and *all* messages, not just the message(s).

[Sanjay_6]

Hi David,

Try this commands,

CD /etc
tar cvfb /dev/rmt/0m 20 /etc
tar tvfb /dev/rmt/0m 20

cd /tmp
mkdir test_dir
cd test_dir
tar xvfb /dev/rmt/0m 20

Hope this helps.

Regds

[Darrell Allen]

Hi David,

Please post the output to these:

ioscan -fnC tape (use the n option so you get the device path)

ls -l on every device path for the tape drive from your ioscan

Hopefully ioscan shows /dev/rmt/0m since that's the default for tar. And hopefully ls -l /dev/rmt/0m looks okay.

Thanks,
Darrell

[Volker Borowski]

Another approach:

Insert a DDS3 tape into the drive. If it is not automatically ejected, you may have a jumper problem (MRS ist set to inactive).
In this case, it may be possible, that the drive does not switch to the appropriate density when writing.

Are you trying DDS1 or DDS2 media ?

Do not know if this helps
Volker

[Volker Borowski]

Darell has a point,

in addition check environment variable $TAPE, in case the default is overridden !

Volker

[Carlos Fernandez Riera]

Hi David:

For i have read you tape drive is broken. It can write but have lots os problem while reading.

First check this thread:

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x298bee3e323bd5118fef0090279cd0f9,00.html

There there is a source of a full DDS diagnostic source. I had to wite it because lots of problems with four DDS2 i an K server. Due to intensive use of this drivers i had to change a drive each two months.

Compile and run it. Run tar again and run it again. Search for differences. It must losts of raw retrys and Error correction algorithm.


2- Change method of read. Now use
dd if=/dev/rmt/0m of=/dev/null bs=10k

Its very posible that dd return I/O error.

Change driver and use new tapes, because due to errors tape sould be degradated.

[Frank Slootweg]

I wrote, but got no answer to:

> Have you checked the syslog.log and dmesg
> (1M) output for errors? Note that both are
> buffered, so you may see the errors only
> some time after they occured.

I agree with Carlos that your DDS drive is probably defective. Contrary to what their name implies, DDS drives are effectively analog devices which work like video-recorders (i.e. using the heliscan technique). While *normally* very reliable, they can develop problems, especially alignment problems, resulting in intermittent behaviour like you are experiencing. So my advice is to have the drive checked/repaired/replaced.

[Paula J Frazer-Campbell]

David

Errors occuring at different places ???
Could be software, but unlikly.

Go for the weakest link -

Change the Dat drive , yes I know that you have done it once but how certain are you that it is 100% OK.

Paula

[David Preston_4]

No syslog entries for errors,
i created this box with ignite from my current system (another K) wouldn't it have the same patch level as the ignite server?

Dave
Looks like I am back at the drive.

[David Preston_4]

You Guys have been a great help, I swapped the drive with my production server and it worked just fine, so it appears that it is not the hardware after all.

Are there drivers or other pieces I need to check out in addition to tryiong to get the extension patch bundle?

Dave