Buslogic SCSI, 1.3.5x kernels broken?

Buslogic SCSI, 1.3.5x kernels broken?

Post by Michael Pee » Mon, 29 Jan 1996 04:00:00



Greetings, fellow Linuxers,

        Not too long ago, I posted about having fatal problems using ext2fs on a
SCSI drive connected to a Buslogic SCSI card.  I have been monitoring this group
since then and I have neither recieved a reply, nor noticed anyone with similar
problems.  I was wondering if anyone else out there has had problems using Buslogic
cards - or is it just me?  Should I risk trying the newer kernels or wait for the
next stable release?

        Basically, to recap:

        I downloaded and compiled the latest bleeding-edge kernel, at the time
1.3.52.  When rebooting, I noticed that the Buslogic driver had changed, as it
looked different during bootup.  (Other than that I took little notice.)  But after
a few hours of normal use, X locked up.  I was unable to get out of X, nor was it
paying attention to the three-finger-solute. (X's or Linux's.)

        When I rebooted, Linux complained about having to check the drive.  I
expected that.  I chalked it up to life on the egde.  But then fsck said that the
magic number was corrupt.  It didn't even know what kind of drive it was!  So I ran
fsck manually and told it what kind of drive it was and the inode density.  It
complaned vehemently for about 3 hours - AND THEN GAVE UP!  The drive was still
kaput.

        So I reformatted the drive and restored from tape.  Then, as an experiment,
I tried doing a shutdown.  During the shutdown the kernel froze.  It got as far as
"Sending all processes the TERM signal..." but never got to "System halted."

        With a feeling of dread in the pit of my stomach, I hit the reset switch
again.  Once again, fsck didn't even know what kind of drive it was.

        Since then I've been using the old stable kernel with no problems.

        Here's what my syslog said about the whole thing:
--------------------------------------------------------------------------------
kernel: attempt to access beyond end of device
(last message repeated about a billion times)

kernel: EXT2-fs error (device 08:01): ext2_find_entry: bad entry in directory
#157706: rec_len     != 0 - offset=4096, inode=85, red_len=1027, name_len=11
(last message repeated about 5 times)

kernel: EXT2-fs error (device 08:01):ext2_free_blocks: Freeing blocks not in
datazone - block = xxxxxxxxxx, count=1
(this message repeated in some form a billion times)
--------------------------------------------------------------------------------

        Well?  Whattaya think sirs?


 
 
 

Buslogic SCSI, 1.3.5x kernels broken?

Post by Leonard N. Zubkof » Mon, 29 Jan 1996 04:00:00


          Not too long ago, I posted about having fatal problems using
  ext2fs on a SCSI drive connected to a Buslogic SCSI card.  I have been
  monitoring this group since then and I have neither recieved a reply, nor
  noticed anyone with similar problems.  I was wondering if anyone else out
  there has had problems using Buslogic cards - or is it just me?  Should I
  risk trying the newer kernels or wait for the next stable release?

Your previous posting never made it here, I expect.  Please send me via
email the particular model of BusLogic host adapter you have, its hardware
revision (on the physical board itself), and the startup messages the new
driver prints.  Without this information, it's hard to deduce what might be
happening.

With the latest release of the new BusLogic driver, there have been *very*
few problem reports.

                Leonard

 
 
 

Buslogic SCSI, 1.3.5x kernels broken?

Post by Lee Van Dy » Tue, 30 Jan 1996 04:00:00





>      Not too long ago, I posted about having fatal problems using
>  ext2fs on a SCSI drive connected to a Buslogic SCSI card.  I have been
>  monitoring this group since then and I have neither recieved a reply, nor
>  noticed anyone with similar problems.  I was wondering if anyone else out
>  there has had problems using Buslogic cards - or is it just me?  Should I
>  risk trying the newer kernels or wait for the next stable release?

I've experienced similiar problems with 1.3.57, 1.3.58,
of which where compiled according to the README on a
486/66 VLB Motherboard
Buslogic 542s with the latest bios upgrades as of 11/95. (isa)
Seagate 2.7 Gig scsi-drive (Model ?)
20 Mb Ram
vlb ide
scsi tape
gcc 2.7.1
make 3.74

The problem occurred when I tried to copy from one partion to another.
i.e.
/dev/sda3 -> /dev/sda4
mounted as "/" and "/usr2".

1st the bin directory disappeared, and upon rebooting corruption
messages. I also tried using the e2fsck -b 8192 option with
no luck. (1.3.57)

2nd time, similar results (1.3.58)

3rd time the soft link, /usr/src/linux, which did point properly
to the distribution directory, got converted into a block device
and the file systems was corrupted again. (After trying to copy
files from Partion 3 to 4).

--
Lee Van*            Senior Software Eng.       714-852-1000(voice)    

2181 Dupont Dr.        
Irvine, Ca 92715, USA   http://www.veryComputer.com/          

 
 
 

Buslogic SCSI, 1.3.5x kernels broken?

Post by Leonard N. Zubkof » Tue, 30 Jan 1996 04:00:00



  I've experienced similiar problems with 1.3.57, 1.3.58,
  of which where compiled according to the README on a
  486/66 VLB Motherboard
  Buslogic 542s with the latest bios upgrades as of 11/95. (isa)
  Seagate 2.7 Gig scsi-drive (Model ?)
  20 Mb Ram
  vlb ide
  scsi tape
  gcc 2.7.1
  make 3.74

  The problem occurred when I tried to copy from one partion to another.
  i.e.
  /dev/sda3 -> /dev/sda4
  mounted as "/" and "/usr2".

  1st the bin directory disappeared, and upon rebooting corruption
  messages. I also tried using the e2fsck -b 8192 option with
  no luck. (1.3.57)

  2nd time, similar results (1.3.58)

  3rd time the soft link, /usr/src/linux, which did point properly
  to the distribution directory, got converted into a block device
  and the file systems was corrupted again. (After trying to copy
  files from Partion 3 to 4).

Have you tried the same hardware configuration with 1.2.13?  I see mixed
reports on the stability of the recent 1.3.5[789] kernels on the
linux-kernel mailing list, and there have been a lot of changes to the
buffer/page cache recently.

                Leonard

 
 
 

Buslogic SCSI, 1.3.5x kernels broken?

Post by Piercarlo Gran » Thu, 01 Feb 1996 04:00:00






>>>> Not too long ago, I posted about having fatal problems using ext2fs
>>>> on a SCSI drive connected to a Buslogic SCSI card.  I have been
>>>> monitoring this group since then and I have neither recieved a
>>>> reply, nor noticed anyone with similar problems.  I was wondering
>>>> if anyone else out there has had problems using Buslogic cards - or
>>>> is it just me?  Should I risk trying the newer kernels or wait for
>>>> the next stable release?

lee> I've experienced similiar problems with 1.3.57, 1.3.58,
lee> of which where compiled according to the README [ ... ]

Odds are that this has got nothing to do with the Buslogic driver. the
1.3.x releases are ALPHA and contain totally untested code (some of
which often does not even compile successfully), that can easily thrash
hard disks.

You should be using an ALPHA kernel only if you are prepared to debug it
yourself... For otherwise it is pretty hard to know which particular new
bit of code might be responsible.

 
 
 

Buslogic SCSI, 1.3.5x kernels broken?

Post by Lee Van Dy » Thu, 08 Feb 1996 04:00:00


Hi, I've found a solution to the previous buslogic/1.3.5x problem
(486-66 vlb, B.L. 542 isa, seagate 2.7 gig).

I am now using an ide drive for the system partition and everything
works fine up to 1.3.59. This system is not networked.

However, on a different system:

Pentium 133
Micronics M54i PCI
Buslogic 946c PCI
3com 3c590
Conner/Seagate 4 Gigs
gcc 2.7.1
make 3.74

networking is broken. 1.3.27 works fine.
Anyone else having similar problems.

--
Lee Van*            Senior Software Eng.       714-852-1000(voice)    

2181 Dupont Dr.        
Irvine, Ca 92715, USA   http://www.veryComputer.com/          

 
 
 

1. Buslogic SCSI, 1.3.5x kernels broken?

Greetings, fellow Linuxers,

        Not too long ago, I posted about having fatal problems using ext2fs on a SCSI
drive connected to a Buslogic SCSI card.  I have been monitoring this group since
then and I have neither recieved a reply, nor noticed anyone with similar problems.
I was wondering if anyone else out there has had problems using Buslogic cards - or
is it just me?  Should I risk trying the newer kernels or wait for the next stable
release?

        Basically, to recap:

        I downloaded and compiled the latest bleeding-edge kernel, at the time 1.3.52.
When rebooting, I noticed that the Buslogic driver had changed, as it looked
different during bootup.  (Other than that I took little notice.)  But
after a few hours of normal use, X locked up.  I was unable to get out of X, nor
was it paying attention to the three-finger-solute. (X's or Linux's.)
        When I rebooted, Linux complained about having to check the drive.  I expected
that.  I chalked it up to life on the egde.  But then fsck said that the magic
number oas corrupt.  It didn't even know what kind of drive it was!  So I ran fsck
manually and told it what kind of drive it was and the inode density.  It complaned
vehemently for about 3 hours - AND THEN GAVE UP!  The drive was still kaput.
        So I reformatted the drive and restored from tape.  Then, as an experiment,
I tried doing a shutdown.  During the shutdown the kernel froze.  It got as far as
"Sending all processes the TERM signal..." but never got to "System halted."
        With a feeling of dread in the pit of my stomach, I hit the reset switch again.
Once again, fsck didn't even know what kind of drive it was.

        Since then I've been using the old stable kernel with no problems.

        Well?  Whatta think sirs?


2. print # while copying each file one directory to another.

3. ppp is broke on latest 1.3.5x kernels

4. Need PKZip compatible program for SUNOS4.1.3

5. kernel bdflush problems 1.3.5x

6. Large IDE drives ? 8+Gig

7. Where is Kernel 1.3.5X?

8. System usage statistics program

9. Voxware 1.3.01 with kernel 1.3.5x how?

10. Problems with DIP and kernel 1.3.5x

11. 1.3.5x and bdflush problem

12. upgrade 1.3.24 to 1.3.5x???

13. 1 ide, 2 esdi (WD1007) - 1.3.5x fails ?