Bug 105612

Summary: Problem with parted or grub
Product: [openSUSE] SUSE LINUX 10.0 Reporter: Andreas Mantke <maand>
Component: YaST2Assignee: Thomas Fehr <fehr>
Status: RESOLVED FIXED QA Contact: Klaus Kämpf <kkaempf>
Severity: Critical    
Priority: P5 - None    
Version: Beta 1   
Target Milestone: Beta 3   
Hardware: 64bit   
OS: SUSE Other   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Attachments: yast2log of the installation
the part of /var/log/messages from the start of SuSE 9.3
part with former hdscan in /var/log/messages
/var/log/boot.msg

Description Andreas Mantke 2005-08-18 15:58:06 UTC
I tried yesterday to install OpenSuSE10-Beta on my box. First I  
found a problem with parted. This program was not able to read the  
partition-table of my pc. I am not able to put a new partition on  
hdb. I had to do this with Yast/parted from 9.3 (on the same box). 
Later - after the first installation of 10.0 on the box - there are  
problems with grub (from 10.0). I am not able to start 9.3 again  
and run into a kernel-panic short after the start (root-partition  
of 9.3). I tried to fix this problem with the repair-modus of 9.3.  
But there was no change. I run again in this trouble. 
Only a full update of 9.3 with installing grub of this distri fixed  
the problem. 
I think, OpenSuSE had made some changes on the root-partition of 9.3  
or there are a problem with grub (10.0). 
My box is a AMD64-3000 with two harddisks, 512MB Memory, Asus  
K8N-Board.
Comment 1 Andreas Mantke 2005-08-18 16:02:22 UTC
Created attachment 46568 [details]
yast2log of the installation
Comment 2 Thomas Fehr 2005-08-18 16:12:12 UTC
The problelm with your disk /dev/hdb is that /proc/partitions says there are
partitions up to /dev/hdb21 while parted only displays /dev/hdb17. Do you have
any idea why this is the case?
Comment 3 Andreas Mantke 2005-08-18 17:20:34 UTC
No idea. I see under 9.3 only /dev/hdb17 (OpenSuSE 10) as last entry. 
Comment 4 Thomas Fehr 2005-08-18 17:25:05 UTC
What do you see in /proc/partitions in 9.3.
Comment 5 Andreas Mantke 2005-08-18 17:41:41 UTC
cat /proc/partitions 
major minor  #blocks  name 
 
   3     0   30037392 hda 
   3     1    6144831 hda1 
   3     2    1052257 hda2 
   3     3     313267 hda3 
   3     4          1 hda4 
   3     5    4096543 hda5 
   3     6    4096543 hda6 
   3     7     530113 hda7 
   3     8   10490413 hda8 
   3     9    3309358 hda9 
   3    64  156290904 hdb 
   3    65   10483200 hdb1 
   3    66          1 hdb2 
   3    69   10490413 hdb5 
   3    70   10490413 hdb6 
   3    71   10490413 hdb7 
   3    72   10490413 hdb8 
   3    73     546178 hdb9 
   3    74   10490413 hdb10 
   3    75   10482381 hdb11 
   3    76    1044193 hdb12 
   3    77     313236 hdb13 
   3    78   10490413 hdb14 
   3    79   10490413 hdb15 
   3    80   10490413 hdb16 
   3    81   31455238 hdb17 
   3    82    4507776 hdb18 
   3    83     608832 hdb19 
   3    84   10475136 hdb20 
   3    85    5346432 hdb21 
I don't know why there are 21 partitions. 
Comment 6 Thomas Fehr 2005-08-18 17:49:52 UTC
If you do ot know that I certainly do not know either.
Maybe you have some BSD systems installed in some partitions these might cause
such things. Maybe I can see something if you attach /var/log/messages the
kernel normally tells which partitions it find why.

Whatever caused this, it is the reason that YaST2 warns you about changeing
anything on the partitioning of /dev/hdb.
Comment 7 Andreas Mantke 2005-08-18 19:07:31 UTC
Created attachment 46597 [details]
the part of /var/log/messages from the start of SuSE 9.3
Comment 8 Andreas Mantke 2005-08-19 19:38:46 UTC
Hello, 
are enough information in the part of the log? Should I put some more 
information to this Issue? 
Comment 9 Thomas Fehr 2005-08-19 22:54:25 UTC
No the interesting part is where the partition table is scanned it looks similar 
to the following:

Mar 20 00:15:53 fermat kernel: SCSI device sda: 312581808 512-byte hdwr sectors
(160042 MB)
Mar 20 00:15:53 fermat kernel: sda: assuming drive cache: write through
Mar 20 00:15:53 fermat kernel:  sda: sda1 sda2
Mar 20 00:15:53 fermat kernel: Attached scsi disk sda at scsi1, channel 0, id 0,
lun 0
Mar 20 00:15:53 fermat kernel: Attached scsi generic sg1 at scsi1, channel 0, id
0, lun 
Comment 10 Andreas Mantke 2005-08-20 07:55:24 UTC
Created attachment 46755 [details]
part with former hdscan in /var/log/messages

I could not found a text like yours in /var/log/messages. My darddrive is not a
scsci-modell, so the messages will be not the same. I hope this part of my log
is the right. The full log is in my opinion to big for this issue.
Comment 11 Thomas Fehr 2005-08-20 09:49:11 UTC
Unfortunately the partition scan is still not included.
If in doubt simply attach /var/log/boot.msg this contains all hardware related
info from last system boot.
Comment 12 Andreas Mantke 2005-08-20 11:48:21 UTC
Created attachment 46757 [details]
/var/log/boot.msg

I think the problem came from the solaris10-installation in hdb1. I see in the
log, that there are 3 or 4 partitions are created in hdb1.
Comment 13 Thomas Fehr 2005-08-22 10:42:07 UTC
You are right the bsd slices of istalled solaris are respinsible for the
added partitions. I changed the consistency check to be less struct when
checking. Please retry installation with beta#3.