IBM 1GByte disk (DPES 31080) crashes Slackware with 1.2.13 (repost)

IBM 1GByte disk (DPES 31080) crashes Slackware with 1.2.13 (repost)

Post by Ulrich Thoen » Fri, 22 Dec 1995 04:00:00



Hello from Heidelberg.

In some sense my problem might touch both groups.
So please don't be bad with me if I am wrong in info.grass.user

Grass, a very good GIS system running also on linux requires handling
with sometimes big datasets. (A gziped file with 24MBytes results in
some 68MBytes unziped). When I tried to handle this files regardless
they were gziped or not I run into problems meaning that my whole
Linux System broke down. The only way out of it was to do a hardware
reset. (I know, shouldn't be done with linux).
WHat I mean with handeling is:

Copying or moving from one directory to another regardless these
directories are on the same partition / disk or not.

Un-taring a big tar file (tar -xf xxx.tar)

gezping a large .gz file (gzip -d xxx.tar.gz)

Especially in the last case the process stoped an some point resulting
in a partial ungziped file as I found out after reboot and
reorganizing the filesystem with a fsck during bootup.
Even when mounting the big 500MByte partition I some times get a
system crash.

I tried to work on both a primary partition or a logical partition on
a extended partition. The filesystem I used in both cases was ext2.
The size of the partition was in both cases about 500MByte of a 1090
Mbyte disk. (IBM DPES 31080)

The interesting thing is that I did not had similar problems when
worked on a 400MByte Partition of the same disk under DOS-Windows.
I actually did not try to handel that big files. The biggest I
handeled was som 3.5MByte. This worked without problems.

To make the storry complete I should say, that I succeded sometimes to
handel the big files just directly after switch on of my computer and
boot of Linux. I very recently had a problem with one of my power
supply kabels not giving the right contact, resulting in  some 9.8
Volts instead of 12.00 Volts. I was able to fix this problem now and
do have stable 12.0 Volts all the time.
In addition when I do some formating or some fdisk stuff I get the
following warning:

(...)
The number of cylinders for this disk is set to 1034.
This is larger than 1024, and may cause problems with some software.
(...)
----------------------------------------------------------------------
Here you'll find my total organization of my four disks.

IDE drive:
/dev/hda1           1       1     723  151812+   6  DOS 16-bit >=32M
/dev/hda2         732     732     965   49140   83  Linux native
/dev/hda3   *     724     724     731    1680    a  OS/2 Boot Manager
/dev/hda4         966     966     989    5040   82  Linux swap

1st SCSI drive:
/dev/sda1           1       1     150  153584    6  DOS 16-bit >=32M
/dev/sda2         151     151     206   57344   83  Linux native

2nd SCSI drive:
/dev/sdb1           1       1     107  109552    6  DOS 16-bit >=32M
/dev/sdb2         108     108     346  244736    5  Extended
/dev/sdb3         347     347     507  164864   83  Linux native
/dev/sdb5   *     108     108     246  142320    7  OS/2 HPFS
/dev/sdb6   *     247     247     346  102384    6  DOS 16-bit >=32M

3rd SCSI drive:
/dev/sdc1           1       1     412  421872    6  DOS 16-bit >=32M
/dev/sdc2         413     413     512  102400   82  Linux swap
/dev/sdc3         514     514    1023  522240    5  Extended
/dev/sdc5         514     514    1023  522224   83  Linux native

/dev/hda1 is the DOS System drive
/dev/sdb3 is the Linux System drive
/dev/sdb5 is the OS/2 System drive

/dev/sdc2 linux swap is as big as 100MByte because of the GIS
application.
----------------------------------------------------------------------
I am running Slackware 2.3 with Kernel 1.2.13

My computer is a i486DX2 66MHz.
Further components are Adaptec SCSI cont. AHA1542CF,
VL bus controler,
3rd and 4th serial high speed port
4speed cdrom plextor, SCSI
1 Speed cdrom Mitsumi (the very old one)
Soundblaster pro,
Tsengslab ET4000W32 graphic card
Connor QIC Floppy streamer

This got a long posting. However it should be complete.
I would be very thankfull for any help.

Thanks Ulli


 
 
 

IBM 1GByte disk (DPES 31080) crashes Slackware with 1.2.13 (repost)

Post by Andries Brouw » Sat, 23 Dec 1995 04:00:00


: Grass, a very good GIS system running also on linux requires handling
: with sometimes big datasets. (A gziped file with 24MBytes results in
: some 68MBytes unziped). When I tried to handle this files regardless
: they were gziped or not I run into problems meaning that my whole
: Linux System broke down. The only way out of it was to do a hardware
: reset.

Big files are no problem under Linux - I often use 500MB files.
You must have a bad setup or bad hardware.
"my whole Linux System broke down" is not precise enough to guess
what your problem might be.
[I hope you know how to properly shut down your machine.]

 
 
 

IBM 1GByte disk (DPES 31080) crashes Slackware with 1.2.13 (repost)

Post by Don Dettk » Sat, 23 Dec 1995 04:00:00


<snipped>  ( complaints of Linux crashing on large I/Os )

Quote:>To make the storry complete I should say, that I succeded sometimes to
>handel the big files just directly after switch on of my computer and
>boot of Linux. I very recently had a problem with one of my power
>supply kabels not giving the right contact, resulting in  some 9.8
>Volts instead of 12.00 Volts. I was able to fix this problem now and
>do have stable 12.0 Volts all the time.

<snip>

Quote:>Here you'll find my total organization of my four disks.
>IDE drive:
>/dev/hda1           1       1     723  151812+   6  DOS 16-bit >=32M
>/dev/hda2         732     732     965   49140   83  Linux native
>/dev/hda3   *     724     724     731    1680    a  OS/2 Boot Manager
>/dev/hda4         966     966     989    5040   82  Linux swap
>1st SCSI drive:
>/dev/sda1           1       1     150  153584    6  DOS 16-bit >=32M
>/dev/sda2         151     151     206   57344   83  Linux native
>2nd SCSI drive:
>/dev/sdb1           1       1     107  109552    6  DOS 16-bit >=32M
>/dev/sdb2         108     108     346  244736    5  Extended
>/dev/sdb3         347     347     507  164864   83  Linux native
>/dev/sdb5   *     108     108     246  142320    7  OS/2 HPFS
>/dev/sdb6   *     247     247     346  102384    6  DOS 16-bit >=32M
>3rd SCSI drive:
>/dev/sdc1           1       1     412  421872    6  DOS 16-bit >=32M
>/dev/sdc2         413     413     512  102400   82  Linux swap
>/dev/sdc3         514     514    1023  522240    5  Extended
>/dev/sdc5         514     514    1023  522224   83  Linux native
>/dev/hda1 is the DOS System drive
>/dev/sdb3 is the Linux System drive
>/dev/sdb5 is the OS/2 System drive

<snip>

Quote:>I am running Slackware 2.3 with Kernel 1.2.13
>My computer is a i486DX2 66MHz.
>Further components are Adaptec SCSI cont. AHA1542CF,
>VL bus controler,
>3rd and 4th serial high speed port
>4speed cdrom plextor, SCSI
>1 Speed cdrom Mitsumi (the very old one)
>Soundblaster pro,
>Tsengslab ET4000W32 graphic card
>Connor QIC Floppy streamer

<snip>

I have no problems copying or untarring files larger than 100MB on my
DPES31080.  I heard a rumor a while back that it didn't work with Tagged
Command Queuing.  However, I cannot confirm or deny that, as the driver
for my FD1680 doesn't implement it.  Anyway, if you ARE using Tagged
Command Queuing, you might see if disabling it helps.

---
Don Dettke,  Software Engineering Consultant

 
 
 

IBM 1GByte disk (DPES 31080) crashes Slackware with 1.2.13 (repost)

Post by Ulrich Thoen » Mon, 25 Dec 1995 04:00:00




>: Grass, a very good GIS system running also on linux requires handling
>: with sometimes big datasets. (A gziped file with 24MBytes results in
>: some 68MBytes unziped). When I tried to handle this files regardless
>: they were gziped or not I run into problems meaning that my whole
>: Linux System broke down. The only way out of it was to do a hardware
>: reset.

>Big files are no problem under Linux - I often use 500MB files.
>You must have a bad setup or bad hardware.
>"my whole Linux System broke down" is not precise enough to guess
>what your problem might be.
>[I hope you know how to properly shut down your machine.]

That's for sure how to shut down the machine.
...broke down means:

During a large I/O action, like moving or copying as well untaring or
unzipping larg files (20-60Mbyte) on that partition. I can not do
anything with my system anymore. No kyestroke, no mouse movement.
Absolutly nothing. Shut down with CTRL + ALT +DEL brings NO effect.
The ONLY possibility is to push the reset bittom.
This leads to an automatic fsck on the specified partition when
rebooting.

Ulli

 
 
 

1. Adaptec 1515 with IBM DPES 31080

I've just bought a new IBM DPES 31080 (SCSI) 1GB harddisk,
which I connected to my Adaptec 1515 Controller.
I booted from my old IDE drives and used the
"afdisk" utility, which was supplied with the controller,
to partition the new disk. I installed DOS 6.0 on the first
two partitions of the disk and everything worked
fine until I tried to install linux on the last three
partitions (one of them for swapping).
I used linux fdisk to change the  partiontions to linux format,
and loaded the data from my old harddisk.
Unfortunately fdisk -v reports several errors
(e.g. overlaying sectors and problems with the disk
geometry) and when I tried to install lilo it
reported problems as well. After adding "linear" to the
lilo.conf lilo finally worked, but it is very slow,
when it starts linux.
When I tried to partition the disk using the linux fdisk,
everthing is fine for linux, only that DOS then reports
problems when it tries to format its own partions (which
I set to MSDOS format under linux fdisk) and cant find its own
second partition.

So: whatever way I try, either DOS or Linux seem to have some
trouble with the partitions on my harddisk.

Does anyone know these problems? Is it the controller or linux
or DOS?


2. Linux Deployment Tools

3. Adaptec AHA-1742 with IBM DPES-31080 harddisk

4. PPP Dialup to NT4.0

5. Adaptec 1542 + IBM DPES 31080 + Kernel 1.2.11 Problem

6. Sendmail

7. IBM 1Gb DPES-31080 on Solaris 2.5

8. man failure, what gives?

9. SOLUTION: Adaptec 1542CF, IBM DPES 31080 and other SCSI devices

10. Adaptec 1542CF, IBM DPES-31080, SCSI Probs

11. Asus SC200 and IMB DPES-31080, Linux-1.3.50

12. Netscape/Java crashes with Slackware 1.2.13

13. Slackware 3.0 'ascsi' boot disk is 1.3.18 (I want 1.2.13)