Hey All:
I'm posting this summary quite late; my apologies. I got a lot of enlightening
responses, unfortunately none of them cured my disk. I suspect the problem
must be internal and probably needs to be replaced. In any case, let me share
the great feedback I received. I've interspersed my own commentary. The
original question is at the bottom.
--------------- BEGIN REPLIES -------------------
-------------------------------------------------------------------------------
>From: strombrg@hydra.acs.uci.edu(Dan Stromberg)
>Try relabling the disk first.
>
>Then try this:
>
> - apparently sometimes the block number that is reported bad by the
> kernel is off, due to scatter/gather-ish stuff. This might explain
> the reports of "repair" failing.
> - one can apparently translate the block # to track/cyl/head with show,
> then analyze a few surrounding tracks (being careful to answer "no" to
> "repair defective blocks?") to find the -really- bad track/cyl/head,
> (by specifying that analyze transfer data one block at a time!) and
> then use "show" again to convert back to the true bad block number
> - This is supposed to yield the actual bad block (or blocks, of course)
> - from there, one can "repair" that block or...
> - find that file, and just stop using it for the time being
> - that "disk block number" can be converted to a "partition
> block number", based on the sizes of the partitions
> - an inode can then be found with "icheck", if the block is
> allocated to a file
> - then directory entries corresponding to the inode can be found
> with "ncheck" or "find -inum"
>
>If neither work, try dd'ing the raw filesystem, see if anything's
>readable. If it is, you may be able to recover some of your data
>(especially ascii) without that much trouble. If you need more,
>consider a data recovery service.
>
>Then format. :)
format can see the disk, display the partition. It can analyze every sector
and cylinder. It will (apparently) allow me to repartition UNTIL I TRY TO
WRITE IT TO DISK WITH THE LABEL COMMAND. Then I get the following error:
> No sense error during write
> ASC: 0x0 ASCQ: 0x0
> Warning: error writing primary label.
> No sense error during write
> ASC: 0x0 ASCQ: 0x0
repeating incessantly! I don't know what to make of this. So, I cannot
repartition, repair, or format. I'M STILL OPEN TO SUGGESTION!
--------------------------------------------------------------------------------
>From: blymn@mulga.awadi.com.AU (Brett Lymn)
>The format analyze/repair option does not function on SCSI devices, it
>is there for bad block maintenance on the old ST506 drives. To fix a
>bad block on a SCSI device you have to reformat the drive.
Hmmmm, well, format didn't complain when I chose to analyze the disk, it
simply didn't find any errors.
--------------------------------------------------------------------------------
>From: David Fetrow <fetrow@biostat.washington.edu>
> I had one. Reformatting worked once but not the second time it failed.
>They should have 5 year warranties.
This sounds like the unfortunate winner.
--------------------------------------------------------------------------------
>From: eckhard@ts.go.dlr.de (Eckhard Rueggeberg)
>Perhaps you should leave out the 0th cylinder when you repartition the disk ...
Great idea, I'm sure... if only I could repartition.
--------------------------------------------------------------------------------
>From: Luiz Claudio Schara Magalhaes <schara@exu.inf.puc-rio.br>
>Before taking extreme measures, try disconnecting and connecting the SCSI
>cables. Sometimes they are loose, and you will see all kind of strange errors.
Words of wisdom, alas, of no avail.
----------------- END OF REPLIES -------------------
--------------------------------------------------------------------------------
ORIGINAL QUESTION:
>I have a Sun SPARC IPX with a failed disk. The disk is in a box
>labeled PARITY SCSI Bootbox which contains two scsi devices, the hard
>disk in question and a tape drive (which works and we don't care
>about). The hard disk is identified as a CDC Wren IV according to
>FORMAT:
>
>>AVAILABLE DISK SELECTIONS:
>> 0. sd0 at esp0 slave 24
>> sd0: <SUN0207 cyl 1254 alt 2 hd 9 sec 36>
>> 1. sd3 at esp0 slave 0
>> sd3: <CDC Wren IV 94171-344 cyl 2196 alt 2 hd 7 sec 41
>>Specify disk (enter its number): 1
>>selecting sd3: <CDC Wren IV 94171-344
>>[disk formatted, defect list found]
>
>I believe this is correct. I don't seem to have any documentation for
>this device. The error message I get a boot time, when it attempts to
>mount this device is as follows:
>
>>sd3c: Vendor 'CDC' error code:0x11
>>sd3c: Target 0 now Synchronous at 4.0 mb/s max transmit rate
>>sd3c: Error for command read
>>sd3c: Error level retryable
>>sd3c: Block 16, Absolute block: 16
>>sd3c: Sense key: Media error
>
>I've tried to use FORMAT's analyze and repair facility's to no avail.
>Analyze doesn't find any bad blocks on or near block 16 and the repair
>option fails. My next step is to try to reformat the whole disk. I
>thought I'd ask the Sun Manager Superpowers for advice before I take
>this next horrible step. Do I, in fact, have a bad block or is it
>something else? Does anyone else out there have one of these odd-ball
>devices?
>
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:08:02 CDT