Very Computer

Hi !

        By trying to retrieve too much (Usenet) articles, my swap
partition blew off (100 Mb)...

        This partition used to be hda7, and when I fdisk /dev/hda, it
gaves me strange numbers for this partition, so I deleted it. Now,
when I try to fdisk hda, I have:  

Using /dev/hda as default device!

Unable to read /dev/hda

        When I boot, I have these messages when the kernel tries to
check the disk:

Jul  8 13:39:25 entropie kernel: Partition check:
Jul  8 13:39:25 entropie kernel:  hda:Started kswapd v 1.4.2.2
Jul  8 13:39:25 entropie kernel:  hda1 hda2 < hda5 hda6hda: read_intr:
status=0x59 { DriveReady SeekComplete DataRequest Error }
Jul  8 13:39:25 entropie kernel: hda: read_intr: error=0x10 {
SectorIdNotFound }, LBAsect=36933219, sector=0
Jul  8 13:39:25 entropie kernel: hda: read_intr: status=0x59 {
DriveReady SeekComplete DataRequest Error }
Jul  8 13:39:25 entropie kernel: hda: read_intr: error=0x10 {
SectorIdNotFound }, LBAsect=36933219, sector=0
Jul  8 13:39:25 entropie kernel: hda: read_intr: status=0x59 {
DriveReady SeekComplete DataRequest Error }
Jul  8 13:39:25 entropie kernel: hda: read_intr: error=0x10 {
SectorIdNotFound }, LBAsect=36933219, sector=0
Jul  8 13:39:25 entropie kernel: hda: read_intr: status=0x59 {
DriveReady SeekComplete DataRequest Error }
Jul  8 13:39:25 entropie kernel: hda: read_intr: error=0x10 {
SectorIdNotFound }, LBAsect=36933219, sector=0
Jul  8 13:39:25 entropie kernel: ide0: reset: success

etc.

        Though the other partitions (hda5 and hda6) are OK, I can't
use fdisk to recreat hda7 or whatever, because I can't use fdisk...
        What can I do, now ?

        Thanks for any help !!!
--

Quote:> Jul  8 13:39:25 entropie kernel:  hda1 hda2 < hda5 hda6hda: read_intr:
> status=0x59 { DriveReady SeekComplete DataRequest Error }
> Jul  8 13:39:25 entropie kernel: hda: read_intr: error=0x10 {
> SectorIdNotFound }, LBAsect=36933219, sector=0
[...]
> Jul  8 13:39:25 entropie kernel: ide0: reset: success

it is perhaps time to write a document!

here is a mix of potential fixes:
 - do not  neglect that it may be a PHYSICAL problem, involving either the
   disk surface, connector, controler... (?)
   you may be sure by using a low-level formater (beware: your data will be
   destroyed)
 - double check all connected devices (master/slave setup)
 - do not activate DMA (with hdparm)
 - use a lower PIO mode (PIO 2 max.)
 - do NOT use the LBA/EIDE functions, configure the drive geometry by hand
 - make only ONE partition per physical drive (?!)
 - disconnect any other EIDE device (especially cheap CD-ROM drives)
 - if you also got "signal 11" errors check http://www.bitwizard.nl/sig11/
 - if you reinstall _always_ keep a file containing a dump of your
   partition table (use "dd") and when the problem arises compare the
   active table with the saved one. update if necessary (warning:
   dangerous)

got many messages like those with an ASUS board dating mid'95. disk (WD
21000) was faulty and I changed it. problems came back, even with a
permissive configuration (no CPU L2 cache, PIO 2 ...) then  I changed also
the motherboard (ASUS TP4N dated jan'96) not a single problem since.

I am affraid that there is a bug in some EIDE controlers or BIOSes. I mean
_another_ bug (not the one related to concurrency or write cache)

check the WEB site of the constructor of your motherboard, BIOS and disk
drive in order to get a test utility, an upgrade software or something...

I also sent copies of this to articles authors who must tackle this
problem.
to all: please keep in touch with me if you have any interesting info!

--
Nat    Linux