Upgrading 11.2->11.3 Corrupted a Different Partition?

Caglow · July 29, 2010, 7:02pm

Yes, I know I said I wouldn’t do anything that could possibly do anything bad to the drive until I receive my external hard drive. I just didn’t expect that upgrading from 11.2 to 11.3 via zypper dup has the potential to do that.

First, the problem: The most important partition of my computer containing all the irreplaceable stuff (images) which I had apparently only partially backed up was corrupted after I upgraded from 11.2 to 11.3.

On my computer, there are 2 drives. sda was the one that had Windows. sdb had an NTFS partition and a few smaller partitions which I used for openSUSE. After doing quite a bit of searching on upgrades corrupting partitions, I didn’t find anything about it doing what I feared which was corrupting the NTFS so I went ahead and upgraded.

The upgrade went along smoothly. No problems that I could tell. The only weird thing was that the window borders disappeared near the end but I assumed that was supposed to happen.

And then I restarted. First thing I noticed was that the background was blue instead of the image I had set (which happened to be on the NTFS partition). I didn’t think much of it at the time figuring it was just a quirk in the upgrade process. I went ahead and decided to reset the background to what it was. And then I realized that the NTFS partition was missing – it wasn’t mounted or visible at all. That was when I started to panic a bit.

I opened up Yast partitioner and everything looked fine except that NTFS partition in question had a little * by it. I went ahead and reset its mount point only to receive an error saying that the filesystem in question doesn’t exist. And sda became sdb and sdb became sda if that changes anything.

I went into Windows expecting for it to do a CHKDISK on bootup for the drive (D: ) but none of that happened. Hoping that everything was fine, I went in and tried to access it only be given an error along the lines of “The drive in D: is not formatted. Do you want to format it?” Of course, I said no but that was when I realized that this was no partition table problem like last time.

I tried restarting but Windows froze and refused to do so for several minutes so forced the computer to shutdown and loaded my copy of GParted LiveCD. It showed a 30GB unrecognized partition, another 30MB one and some unallocated space. TestDisk fixed that. (so it turns out, there was a partition table issue) What was left was what looked OK except the one partition on sda (the original sdb) that I could not afford to lose had an error.

I can’t remember the exact error but it said something like “Are you trying to use a disk as a partition? Are you trying to use /dev/sda as /dev/sda1 or vice versa?” That was either in the error message or in the message I got when I tried to check the disk.

Still trying to get Windows to check the disk, I tried booting from the new sda into Windows. What I got was the Dell Utility program. It said no mouse detected and I couldn’t do anything so I shut it down.

I tried going back into Windows the normal way and got the Dell Utility. I figured that this was an MBR problem after I went back into openSUSE with the new sdb still reading perfectly. Although I couldn’t manage to restore the MBR so I can’t log into Windows. But the other issue is far more important.

So here’s my current plan:

Don’t touch sda1 (the new one)
Gather as much information as possible without touching sda1
Run any tests that can be 100% confirmed to be read-only
Use Clonezilla to clone the entire disk over to an external hard drive (please tell me it can clone corrupted partitions!)
Work off the external hard drive using everything possible
If all fails, clone sda to the external drive again (and I modified the partition on the external drive) and send it to some expert for a few hundred dollars to retrieve data

The only parts I’m not so sure about are the do read-only tests and the disk cloning. I’m pretty sure read-only tests are read-only but look what happened last time I was pretty sure I wouldn’t screw up. The disk cloning because I have no idea if Clonezilla will be able to clone corrupted things.

So I would like some confirmation on the above and some information (other than you should’ve checked your backup) on how to fix things and some possible tests I can run to gather more information.

Thanks for all help received. The fact that I’m willing to spend several hundred dollars or more should tell you how importantly I want this back.

user · July 29, 2010, 7:47pm

On 2010-07-29 19:06, Caglow wrote:
>
> Yes, I know I said I wouldn’t do anything that could possibly do
> anything bad to the drive until I receive my external hard drive. I just
> didn’t expect that upgrading from 11.2 to 11.3 via zypper dup has the
> potential to do that.
>
> First, the problem: The most important partition of my computer
> containing all the irreplaceable stuff (images) which I had apparently
> only partially backed up was corrupted after I upgraded from 11.2 to
> 11.3.
>
> ----
>
> On my computer, there are 2 drives. sda was the one that had Windows.
> sdb had an NTFS partition and a few smaller partitions which I used for
> openSUSE. After doing quite a bit of searching on upgrades corrupting
> partitions, I didn’t find anything about it doing what I feared which
> was corrupting the NTFS so I went ahead and upgraded.

An upgrade does not format any partition, nor change partition layout, grub settings, nothing. An
upgrade from the DVD can do a bit more, but a zypper dup… no, no.

I can’t understand what happened to your system. It has to be something else.

The only thing I can think about is that 11.3 may have problem reading or accessing ntfs partitions.
It is possible to imagine some process writing and corrupting a partition, of course. Ntfs is…
well, dunno. Perhaps not fully understood.

…

> background to what it was. And then I realized that the NTFS partition
> was missing – it wasn’t mounted or visible at all. That was when I
> started to panic a bit.
>
> I opened up Yast partitioner and everything looked fine except that
> NTFS partition in question had a little * by it. I went ahead and reset
> its mount point only to receive an error saying that the filesystem in
> question doesn’t exist. And sda became sdb and sdb became sda if that
> changes anything.

Wow.

> I went into Windows expecting for it to do a CHKDISK on bootup for the

I’m lost with all these manipulations you describe…

> ----
>
> So here’s my current plan:
>
> 1. Don’t touch sda1 (the new one)
> 2. Gather as much information as possible without touching sda1
> 3. Run any tests that can be 100% confirmed to be read-only
> 4. Use Clonezilla to clone the entire disk over to an external hard
> drive (please tell me it can clone corrupted partitions!)

I don’t know about clonezilla. What I know is that is possible to clone an entire disk, or its
partitions one by one, to another disk, or as files to another partition big enough to hold them -
which is my preferred method for attempted recovery.

And yes, you can image a corrupted filesystem, and will simply produce another corrupted filesystem

which is what we want.

If you mean wether you can clone a disk with corrupted partition table, yes, it is also possible to
clone it. The program “dd” will do it. Carefull: dd can as easily destroy any file, partition or
entire disk if you issue the wrong command line.

I have no idea about clonezilla. I would assume it tries to repair things on the fly, dunno.

> 5. Work off the external hard drive using everything possible
> 6. If all fails, clone sda to the external drive again (and I modified
> the partition on the external drive) and send it to some expert for a
> few hundred dollars to retrieve data

That hurts. Sorry.

> The only parts I’m not so sure about are the do read-only tests and the
> disk cloning.

Do the tests on the copy, that’s all. Leave the original not mounted, don’t test anything on it. If
you do the incorrect repair procedure, you still have the original to clone again and repeat the
attempt.

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))

tararpharazon · July 29, 2010, 8:11pm

Yikes!! really sounds bad and messy. I hate to suggest this, but download and burn a copy of Knoppix LiveCD, boot it, and read-only your NTFS and Linux partitions if they exist.
If you can, finish off the backups.

If somehow your /dev/sda became /dev/sdb did you change your Grub menu.lst to boot from the correct partitions? Ditto /etc/fstab is correct?

Menu.lst should show if sda and sdb switched then win should boot off sdb for Windows now, right? Or the uuid names /dev/disk/by-id/yada-yada-part??

I can’t help with Clonezilla, but some of the other clone SW allows you clone using blocks to make an exact copy of the HD, instead of copying the data.
Making an exact clone of a hard disk requires a backup HD of equal or larger size of the target HD.

Caglow · July 29, 2010, 8:27pm

Thanks for the suggestions.

An upgrade does not format any partition, nor change partition layout, grub settings, nothing. An
upgrade from the DVD can do a bit more, but a zypper dup… no, no.

Just what I was thinking which was why I even attempted it. The only other problem that may have occurred was that I grew the affected partition by 1 GB a week ago or so. I could still access it fine from both openSUSE and from Windows. And the data was there two. I could open the files and folders fine too so the data wasn’t corrupted.

> I went into Windows expecting for it to do a CHKDISK on bootup for the

I’m lost with all these manipulations you describe…

Well, normally when Windows boots up, it checks the disks for errors and fixes them. It checked my C: last time I screwed up and solved a few super minor issues. It did nothing for my D: drive (the important one) though.

If you mean wether you can clone a disk with corrupted partition table, yes, it is also possible to
clone it. The program “dd” will do it. Carefull: dd can as easily destroy any file, partition or
entire disk if you issue the wrong command line.

That’s good know. I was just searching around on the internet about how someone corrupted a Vista partition and copied over to 2 different computers. He didn’t seem to have reached a conclusion though.

Do the tests on the copy, that’s all. Leave the original not mounted, don’t test anything on it. If
you do the incorrect repair procedure, you still have the original to clone again and repeat the
attempt.

Okay then, I will do that.

Yikes!! really sounds bad and messy. I hate to suggest this, but download and burn a copy of Knoppix LiveCD, boot it, and read-only your NTFS and Linux partitions if they exist.
If you can, finish off the backups.

Does Knoppix have something that openSUSE doesn’t? openSUSE LiveCD can’t access the partition. My openSUSE 11.3 installation is working fine. That’s what’s strange. I’m even posting this from openSUSE.

If somehow your /dev/sda became /dev/sdb did you change your Grub menu.lst to boot from the correct partitions? Ditto /etc/fstab is correct?

Menu.lst should show if sda and sdb switched then win should boot off sdb for Windows now, right? Or the uuid names /dev/disk/by-id/yada-yada-part??

The upgrade seems to have done that already. Only problem is, I can’t boot into Windows by booting from sdb. And I have no clue about the “full” name of sdb. GRUB seems to just use /dev/sdb. The Windows partition is on sdb2 and the Dell thing is on sdb1 but I see nowhere that sdb1 is ever mentioned.

tararpharazon · July 29, 2010, 8:43pm

Does Knoppix have something that openSUSE doesn’t? openSUSE LiveCD can’t access the partition. My openSUSE 11.3 installation is working fine. That’s what’s strange. I’m even posting this from openSUSE.
Don’t know for sure but when I boot a Knoppix LiveCD there’s the Windows and Linux partitions ready for me in the My Computer file manager. Just a suggestion.

The upgrade seems to have done that already. Only problem is, I can’t boot into Windows by booting from sdb. And I have no clue about the “full” name of sdb. GRUB seems to just use /dev/sdb. The Windows partition is on sdb2 and the Dell thing is on sdb1 but I see nowhere that sdb1 is ever mentioned.
If the upgrade switched it, should it be switched?

Like Carlos E.R., I have trouble following what happened to your installation.
Got any output .png or code (old and new) you can post that might help? /etc/fstab, /boot/grub/menu.lst, /var/log/boot.msg, dmesg ?

Doesn’t the 11.3 upgrade/install process create an install log? How about .rpmnew files?

Please not all output in one huge message, links work.

Caglow · July 29, 2010, 9:05pm

Like Carlos E.R., I have trouble following what happened to your installation.
Got any output .png or code (old and new) you can post that might help? /etc/fstab, /boot/grub/menu.lst, /var/log/boot.msg, dmesg ?

Doesn’t the 11.3 upgrade/install process create an install log? How about .rpmnew files?

There is an install log. Here’s the edited version: UpgradeLog - Google Drive

The original version was similar. The only difference is that there are about 200,000 lines of /] -] ] |] in the middle. I cut that down to around 20. I guess that was something unusual although it could’ve been that my connection was down for a short while. To find where it is, just search for -] accompanied by another 20 or so and you’ll have it.

I actually made a copy of all the things that would get upgraded before I upgraded: upgrade.txt - Google Docs

One strange thing that happened was that the installation got me Wine. I didn’t have Wine installed before – or maybe I did, just with no trace of existence.

Don’t know for sure but when I boot a Knoppix LiveCD there’s the Windows and Linux partitions ready for me in the My Computer file manager. Just a suggestion.

I could try that I guess. Can I be certain that it won’t try to do anything with the partition that may change it?

user · July 29, 2010, 9:24pm

On 2010-07-29 20:35, Caglow wrote:

> Does Knoppix have something that openSUSE doesn’t? openSUSE LiveCD
> can’t access the partition. My openSUSE 11.3 installation is working
> fine. That’s what’s strange. I’m even posting this from openSUSE.

Knoppix is a good live CD system, better at this that oS, I think.

But if you are actually running that 11.3… :-?

Ok, do:


fdisk -l
file -s /dev/sd*

and post it.

The upgrade seems to have done that already. Only problem is, I can’t
boot into Windows by booting from sdb. And I have no clue about the
“full” name of sdb. GRUB seems to just use /dev/sdb.


ls -l /dev/disk/by-uuid/

or

ls -l /dev/disk/by-uuid/  |  grep sdb

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))

user · July 29, 2010, 9:33pm

On 2010-07-29 21:06, Caglow wrote:

>> Doesn’t the 11.3 upgrade/install process create an install log? How
>> about .rpmnew files?
>>
>
> There is an install log. Here’s the edited version:
> http://tinyurl.com/24c2qjh

Please use a pastebin that does not require login.

> The original version was similar. The only difference is that there are
> about 200,000 lines of /] -] ] |] in the middle. I cut that down to
> around 20. I guess that was something unusual although it could’ve been
> that my connection was down for a short while. To find where it is, just
> search for -] accompanied by another 20 or so and you’ll have it.

No, that is (I have not seen your log) a program that is doing a time consuming task, and instead of
not writing anything, it prints a “/”. After an instant, it writes, on top, a “-”. A little bit
later, a “” just on top, overwriting it. What you see is a rotating bar, in text.

In linux, however, when that is written to a file, you get what you saw, because *nix files do not
differentiate a Carriage Return from a Line Feed.

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))

tararpharazon · July 29, 2010, 9:41pm

I’m about 95% certain it won’t mess anything if you don’t change anything. Best I can offer.

Thanks for the links.

To me every thing looks normal except problems found at line 3072 and again at line 3222 with Kate editor.

Perl-Bootloader: 2010-07-28 18:28:10 WARNING: GRUB::GrubDev2UnixDev: No partition found for /dev/disk/by-id/ata-ST380819AS_5MR562CW with 5.

Sadly, I don’t know what those errors mean or what effect they have. There’s other errors, like fonts errors, are normal. And some compile error for memtest.

The /] -] ] |] are animated timing ticks, ie, should be a moving star but it’s displayed slow motion 1 tick per line.

Caglow · July 29, 2010, 9:53pm

> There is an install log. Here’s the edited version:
> Google Docs

Please use a pastebin that does not require login.

Are you sure you have to login? I can access it fine while logged out. Nonetheless, I’m put them on my website:

http://www.caglow.com/upgradelog.txt
http://www.caglow.com/upgrade.txt

First one is the log. Second one is the list of things to upgrade/downgrade/add/remove that I have to confirm before I upgrade.

Here’s the output from fdisk. sdd works fine so won’t worry about it.


Disk /dev/sda: 122.9 GB, 122942324736 bytes
255 heads, 63 sectors/track, 14946 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0xd0f4738c

   Device Boot      Start         End      Blocks   Id  System
/dev/sda1   *           1       11750    94381843+   7  HPFS/NTFS
/dev/sda2           11751       14946    25671870    f  W95 Ext'd (LBA)
/dev/sda5           11751       13969    17824086   83  Linux
/dev/sda6           13970       14883     7341673+  83  Linux
/dev/sda7           14884       14946      506016   82  Linux swap / Solaris

Disk /dev/sdb: 80.0 GB, 80000000000 bytes
255 heads, 63 sectors/track, 9726 cylinders                                                                           
Units = cylinders of 16065 * 512 = 8225280 bytes                                                                      
Sector size (logical/physical): 512 bytes / 512 bytes                                                                 
I/O size (minimum/optimal): 512 bytes / 512 bytes                                                                     
Disk identifier: 0xd0f4738c

   Device Boot      Start         End      Blocks   Id  System
/dev/sdb1   *           1           4       32098+   6  FAT16
/dev/sdb2               5        9726    78091965    7  HPFS/NTFS

Disk /dev/sdc: 255 MB, 255852544 bytes
255 heads, 63 sectors/track, 31 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x000e65cf

   Device Boot      Start         End      Blocks   Id  System
/dev/sdc1   *           1          31      248976    b  W95 FAT32

Disk /dev/sdd: 4040 MB, 4040724480 bytes
125 heads, 62 sectors/track, 1018 cylinders
Units = cylinders of 7750 * 512 = 3968000 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disk identifier: 0x2c6b7369

This doesn't look like a partition table
Probably you selected the wrong device.

   Device Boot      Start         End      Blocks   Id  System
/dev/sdd1   ?      249811      488760   925929529+  68  Unknown
Partition 1 does not end on cylinder boundary.
/dev/sdd2   ?      171637      241182   269488144   79  Unknown
Partition 2 does not end on cylinder boundary.
/dev/sdd3   ?       69548      249981   699181456   53  OnTrack DM6 Aux3
Partition 3 does not end on cylinder boundary.
/dev/sdd4   ?      179952      179955       10668+  49  Unknown
Partition 4 does not end on cylinder boundary.

Partition table entries are not in disk order

file -s /dev/sd*


/dev/sda:  x86 boot sector, Microsoft Windows XP MBR, Serial 0xd0f4738c; partition 1: ID=0x7, active, starthead 1, startsector 63, 188763687 sectors; partition 2: ID=0xf, starthead 254, startsector 188763750, 51343740 sectors, code offset 0xc0
/dev/sda1: x86 boot sector; GRand Unified Bootloader, stage1 version 0x3, LBA flag 0x1, 1st sector stage2 0xb54d2a5, GRUB version 0.97, code offset 0x48, OEM-ID "NTFS    ", sectors/cluster 8, reserved sectors 0, Media descriptor 0xf8, heads 255, hidden sectors 63, dos < 4.0 BootSector (0x80)
/dev/sda2: x86 boot sector; partition 1: ID=0x83, starthead 254, startsector 63, 35648172 sectors; partition 2: ID=0x5, starthead 254, startsector 35648235, 14683410 sectors, code offset 0x7d
/dev/sda5: Linux rev 1.0 ext4 filesystem data (needs journal recovery) (extents) (large files) (huge files)
/dev/sda6: Linux rev 1.0 ext4 filesystem data (needs journal recovery) (extents) (large files) (huge files)
/dev/sda7: Linux/i386 swap file (new style) 1 (4K pages) size 126503 pages
/dev/sdb:  x86 boot sector, Microsoft Windows XP MBR, Serial 0xd0f4738c; partition 1: ID=0x6, active, starthead 1, startsector 63, 64197 sectors; partition 2: ID=0x7, starthead 0, startsector 64260, 156183930 sectors, code offset 0xc0
/dev/sdb1: x86 boot sector, code offset 0x4a, OEM-ID "Dell 8.0", sectors/cluster 4, root entries 512, Media descriptor 0xf8, sectors/FAT 63, heads 255, hidden sectors 63, sectors 64197 (volumes > 32 MB) , serial number 0x7d60118, label: "DellUtility", FAT (16 bit)
/dev/sdb2: x86 boot sector, code offset 0x52, OEM-ID "NTFS    ", sectors/cluster 8, reserved sectors 0, Media descriptor 0xf8, heads 255, hidden sectors 64260, dos < 4.0 BootSector (0x80)
/dev/sdc:  x86 boot sector; partition 1: ID=0xb, active, starthead 1, startsector 63, 497952 sectors, code offset 0x31
/dev/sdc1: x86 boot sector, code offset 0x58, OEM-ID " mkdosfs", Media descriptor 0xf8, heads 255, sectors 497952 (volumes > 32 MB) , FAT (32 bit), sectors/FAT 3831, serial number 0x4c50b54f, label: "           "
/dev/sdd:  x86 boot sector, code offset 0x58, OEM-ID "MSWIN4.1", sectors/cluster 8, reserved sectors 38, Media descriptor 0xf8, heads 255, sectors 7892040 (volumes > 32 MB) , FAT (32 bit), sectors/FAT 7693, reserved3 0x800000, serial number 0xe1a63d6, unlabeled

ls -l /dev/disk/by-uuid/:


total 0
lrwxrwxrwx 1 root root 10 2010-07-29 08:14 07D6-0118 -> ../../sdb1
lrwxrwxrwx 1 root root  9 2010-07-29 10:28 0E1A-63D6 -> ../../sdd
lrwxrwxrwx 1 root root 10 2010-07-29 08:14 4C50-B54F -> ../../sdc1
lrwxrwxrwx 1 root root 10 2010-07-29 08:14 50e73ecd-4720-4b4a-bccc-1def8e77c8af -> ../../sda6
lrwxrwxrwx 1 root root 10 2010-07-29 08:14 681a0446-bf52-4994-a633-fe3fadfccdab -> ../../sda5
lrwxrwxrwx 1 root root 10 2010-07-29 08:14 B280CED680CEA069 -> ../../sdb2
lrwxrwxrwx 1 root root 10 2010-07-29 08:14 c69a2750-2d54-4b46-b936-75084f7b50d6 -> ../../sda7

I am getting ready to burn the Knoppix LiveCD. Also, I would like to be certain about this: I looked around and saw that someone used one and force mounted the partition and it seems to have worked. If I do that, and it doesn’t work, would that damage the partition further? Thanks.

Caglow · July 29, 2010, 10:01pm

Thanks for the links.

To me every thing looks normal except problems found at line 3072 and again at line 3222 with Kate editor.

Perl-Bootloader: 2010-07-28 18:28:10 WARNING: GRUB::GrubDev2UnixDev: No partition found for /dev/disk/by-id/ata-ST380819AS_5MR562CW with 5.

Sadly, I don’t know what those errors mean or what effect they have. There’s other errors, like fonts errors, are normal. And some compile error for memtest.

Oh, apparently, that would be the 100% Windows drive. The other one. And there is no partition 5 on that drive, just 1 and 2.

Caglow · July 29, 2010, 11:04pm

In any case, Knoppix couldn’t “see” the partition either. I did “fix” the Windows not loading issue. Apparently, sometime during all this “chaos”, the boot flag for sda had been set to sda1 instead of the Windows partition which was sda2. Oh, and the weird thing is that Knoppix sees the “correct” sda and sdb drives like openSUSE saw before the upgrade.

tararpharazon · July 29, 2010, 11:31pm

If Knoppix can access sdb2 then I’d back up whatever hidden, configuration and data files that I forgot from the original backup, just in case.
If Knoppix sees all the files on sdb2 thats a start.

Other than sdb1 and sdc1 are bootable partitions it looks to me that sdb2 is still flagged as NTFS.

Can you post to your website the output from :


cat /etc/fstab
#  
cat /boot/grub/menu.lst
#  
cat /var/log/boot.msg

Thanks

Caglow · July 29, 2010, 11:41pm

If Knoppix can access sdb2 then I'd back up whatever hidden, configuration and data files that I forgot from the original backup, just in case.
If Knoppix sees all the files on sdb2 thats a start.

Actually, that partition can be accessed by openSUSE as well. It’s the normal Windows partition. The problem is sda1 (the old sdb1 as seen in oS 11.2 and Knoppix). That, it cannot access. It gets about as far as openSUSE does. Yes, it’s still flagged as an NTFS partition. Problem is, in varying places (including GParted), it gives an error saying “An NTFS filesystem was not found on this partition.”

Here is the log of the result of all the commands: http://www.caglow.com/output.txt

tararpharazon · July 30, 2010, 12:22am

I misread your original post, you’re saying openSUSE and Knoppix read the partition fine but Windows can’t?

If that is true, then I’m about 80% sure you need to mark /dev/sdb1 as a non-boot partition but check-in with Windows gurus in a windows forum to make sure or to determine what needs to be done before you try anything else and harm your system.

Caglow · July 30, 2010, 12:25am

Actually, none of them can read the affected partition. That’s why I’m worrying.

Caglow · July 30, 2010, 12:32am

Now that I look at it, I think you actually misread my post asking whether a force mount could damage the partition. It actually never got to that point in Knoppix. I just saw someone else do it and it worked for them.

user · July 30, 2010, 1:05am

On 2010-07-29 22:06, Caglow wrote:
>
>>
>>> There is an install log. Here’s the edited version:
>>> ‘Google Docs’ (http://tinyurl.com/24c2qjh)
>>
>> Please use a pastebin that does not require login.
>>
>
> Are you sure you have to login? I can access it fine while logged out.

Perhaps you are logged in to gmail on another window, it is the same login.

> Nonetheless, I’m put them on my website:
>
> http://www.caglow.com/upgradelog.txt
> http://www.caglow.com/upgrade.txt

Thanks.

I’ll mention, for other people that don’t have their own site, that we can all upload to a
“pastebin” server. It is explained here: <http://en.wikipedia.org/wiki/Pastebin>. I use the canadian
one: <http://pastebin.ca/>. You paste your text, and you are given a link to it. After a month, it
disappears from the net.

But anyway, I don’t see anything in your logs… or I don’t know what to search for.

> Here’s the output from fdisk. sdd works fine so won’t worry about it.

Ok.

>
> Code:
> --------------------
>
> Disk /dev/sda: 122.9 GB, 122942324736 bytes

> Device Boot Start End Blocks Id System
> /dev/sda1 * 1 11750 94381843+ 7 HPFS/NTFS
> /dev/sda2 11751 14946 25671870 f W95 Ext’d (LBA)
> /dev/sda5 11751 13969 17824086 83 Linux
> /dev/sda6 13970 14883 7341673+ 83 Linux
> /dev/sda7 14884 14946 506016 82 Linux swap / Solaris
>
> Disk /dev/sdb: 80.0 GB, 80000000000 bytes

> Device Boot Start End Blocks Id System
> /dev/sdb1 * 1 4 32098+ 6 FAT16
> /dev/sdb2 5 9726 78091965 7 HPFS/NTFS
>
> Disk /dev/sdc: 255 MB, 255852544 bytes

> Device Boot Start End Blocks Id System
> /dev/sdc1 * 1 31 248976 b W95 FAT32
>
> Disk /dev/sdd: 4040 MB, 4040724480 bytes

> This doesn’t look like a partition table
> Probably you selected the wrong device.
>
> Device Boot Start End Blocks Id System
> /dev/sdd1 ? 249811 488760 925929529+ 68 Unknown
> Partition 1 does not end on cylinder boundary.
> /dev/sdd2 ? 171637 241182 269488144 79 Unknown
> Partition 2 does not end on cylinder boundary.
> /dev/sdd3 ? 69548 249981 699181456 53 OnTrack DM6 Aux3
> Partition 3 does not end on cylinder boundary.
> /dev/sdd4 ? 179952 179955 10668+ 49 Unknown
> Partition 4 does not end on cylinder boundary.

Funny. Perhaps it is not partitioned.

> --------------------
>
>
> file -s /dev/sd*
>
> Code:
> --------------------
>
> /dev/sda: x86 boot sector, Microsoft Windows XP MBR, Serial 0xd0f4738c; partition 1: ID=0x7, active, starthead 1, startsector 63, 188763687 sectors; partition 2: ID=0xf, starthead 254, startsector 188763750, 51343740 sectors, code offset 0xc0
> /dev/sda1: x86 boot sector; GRand Unified Bootloader, stage1 version 0x3, LBA flag 0x1, 1st sector stage2 0xb54d2a5, GRUB version 0.97, code offset 0x48, OEM-ID "NTFS ", sectors/cluster 8, reserved sectors 0, Media descriptor 0xf8, heads 255, hidden sectors 63, dos < 4.0 BootSector (0x80)

So, sda1 is marked bootable, contains grub, and an NTFS filesystem? Is this partition readable? If
this is the one that fails, that is the problem.

Let me see…

In your log I see, when it is installing grub

Root device: /dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part5 (/dev/sdb5) (mounted on / as ext4)

Resume device: /dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part7 (/dev/sdb7)

Root on sdb5? There is no such thing, sdb has only two partitions; it is writing on what fdisk says
it is sda5. If it then writes grub to sda1 it will be destroyed!

I’m not an expert in grub, but IMO, grub can not go on an NTFS partition.

> /dev/sda2: x86 boot sector; partition 1: ID=0x83, starthead 254, startsector 63, 35648172 sectors; partition 2: ID=0x5, starthead 254, startsector 35648235, 14683410 sectors, code offset 0x7d

Not quite right… but perhaps it is. Mine reads like this:

Elessar:~ # file -s /dev/sda4
/dev/sda4: x86 boot sector; partition 1: ID=0x83, starthead 1, startsector 63, 31455207 sectors;
partition 2: ID=0x5, starthead 254, startsector 31455270, 20964825 sectors, extended partition
table, code offset 0x0

I miss the “extended partition table” text in yours. Maybe it is alright, as fdisk sees all
partitions. It does, no?

> /dev/sda5: Linux rev 1.0 ext4 filesystem data (needs journal recovery) (extents) (large files) (huge files)
> /dev/sda6: Linux rev 1.0 ext4 filesystem data (needs journal recovery) (extents) (large files) (huge files)
> /dev/sda7: Linux/i386 swap file (new style) 1 (4K pages) size 126503 pages

Ok.

> /dev/sdb: x86 boot sector, Microsoft Windows XP MBR, Serial 0xd0f4738c; partition 1: ID=0x6, active, starthead 1, startsector 63, 64197 sectors; partition 2: ID=0x7, starthead 0, startsector 64260, 156183930 sectors, code offset 0xc0
> /dev/sdb1: x86 boot sector, code offset 0x4a, OEM-ID “Dell 8.0”, sectors/cluster 4, root entries 512, Media descriptor 0xf8, sectors/FAT 63, heads 255, hidden sectors 63, sectors 64197 (volumes > 32 MB) , serial number 0x7d60118, label: “DellUtility”, FAT (16 bit)
> /dev/sdb2: x86 boot sector, code offset 0x52, OEM-ID "NTFS ", sectors/cluster 8, reserved sectors 0, Media descriptor 0xf8, heads 255, hidden sectors 64260, dos < 4.0 BootSector (0x80)

They seem correct.

I need to see your grub/menu.lst file. I have a suspicion.

> /dev/sdc: x86 boot sector; partition 1: ID=0xb, active, starthead 1, startsector 63, 497952 sectors, code offset 0x31
> /dev/sdc1: x86 boot sector, code offset 0x58, OEM-ID " mkdosfs", Media descriptor 0xf8, heads 255, sectors 497952 (volumes > 32 MB) , FAT (32 bit), sectors/FAT 3831, serial number 0x4c50b54f, label: " "

Correct, I think.

> /dev/sdd: x86 boot sector, code offset 0x58, OEM-ID “MSWIN4.1”, sectors/cluster 8, reserved sectors 38, Media descriptor 0xf8, heads 255, sectors 7892040 (volumes > 32 MB) , FAT (32 bit), sectors/FAT 7693, reserved3 0x800000, serial number 0xe1a63d6, unlabeled

Correct, I think.

>
> --------------------
>
>
> ls -l /dev/disk/by-uuid/:
>
>
> Code:
> --------------------
>
> total 0
> lrwxrwxrwx 1 root root 10 2010-07-29 08:14 07D6-0118 → …/…/sdb1
> lrwxrwxrwx 1 root root 9 2010-07-29 10:28 0E1A-63D6 → …/…/sdd
> lrwxrwxrwx 1 root root 10 2010-07-29 08:14 4C50-B54F → …/…/sdc1
> lrwxrwxrwx 1 root root 10 2010-07-29 08:14 50e73ecd-4720-4b4a-bccc-1def8e77c8af → …/…/sda6
> lrwxrwxrwx 1 root root 10 2010-07-29 08:14 681a0446-bf52-4994-a633-fe3fadfccdab → …/…/sda5
> lrwxrwxrwx 1 root root 10 2010-07-29 08:14 B280CED680CEA069 → …/…/sdb2
> lrwxrwxrwx 1 root root 10 2010-07-29 08:14 c69a2750-2d54-4b46-b936-75084f7b50d6 → …/…/sda7

My mistake. I should have said “by-id”, which I think correspond to those messages in your upgrade log:

Perl-Bootloader: 2010-07-28 18:26:54 WARNING: GRUB::GrubDev2UnixDev: No partition found for

/dev/disk/by-id/ata-ST380819AS_5MR562CW with 5.

And then we can know to what disk it was talking about.

That “GrubDev2UnixDev” function I guess tries to correlate grub device names like hd(0,0) to linux
style device names. If this function gets it wrong… that’s pretty bad.

>
> --------------------
>
>
> I am getting ready to burn the Knoppix LiveCD. Also, I would like to be
> certain about this: I looked around and saw that someone used one and
> force mounted the partition and it seems to have worked. If I do that,
> and it doesn’t work, would that damage the partition further? Thanks.

I don’t know.

If you mean force mounting an ntfs partition that linux is refusing to mount - no, don’t do it. It
usually means that the partition was not properly umounted (in windows) and the log still contains
unfinished transactions that have to be replayed. This can only be done by windows. Otherwise, linux
can only mount it read-only, or forced (r/w), in which case linux erases the log and discards those
changes, which means lost data and perhaps broken files.

I read this somehwere in the ntfs-3g documentation, but from memory, so some details could be wrong.

Please, don’t change any thing. I have a very bad feeling.

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))

Caglow · July 30, 2010, 1:22am

So, sda1 is marked bootable, contains grub, and an NTFS filesystem? Is this partition readable? If
this is the one that fails, that is the problem.

Yes, sdb1 (I’m going to refer to the disks as they originally were so that’s actually sda1) is the one with the problem. GRUB loads fine though so that shouldn’t be the problem. I believe it’s always been on that disk too.

That “GrubDev2UnixDev” function I guess tries to correlate grub device names like hd(0,0) to linux
style device names. If this function gets it wrong… that’s pretty bad.

I believe it did get it wrong. It was trying to look for partition 5 on sda when partition 5 is on sdb.

I need to see your grub/menu.lst file. I have a suspicion.

Here you go:


caglow:/home/caglow # cat /boot/grub/menu.lst
# Modified by YaST2. Last modification on Wed Jul 28 20:58:34 PDT 2010
# THIS FILE WILL BE PARTIALLY OVERWRITTEN by perl-Bootloader
# Configure custom boot parameters for updated kernels in /etc/sysconfig/bootloader

default 0
timeout 8
gfxmenu (hd0,4)/boot/message

###Don't change this comment - YaST2 identifier: Original name: linux###
title Desktop -- openSUSE 11.3 - 2.6.34-12
    root (hd0,4)
    kernel /boot/vmlinuz-2.6.34-12-desktop root=/dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part5 resume=/dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part7 splash=silent quiet showopts vga=0x31a
    initrd /boot/initrd-2.6.34-12-desktop

###Don't change this comment - YaST2 identifier: Original name: failsafe###
title Failsafe -- openSUSE 11.3 - 2.6.34-12
    root (hd0,4)
    kernel /boot/vmlinuz-2.6.34-12-desktop root=/dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part5 showopts apm=off noresume edd=off powersaved=off nohz=off highres=off processor.max_cstate=1 x11failsafe vga=0x31a
    initrd /boot/initrd-2.6.34-12-desktop
                                                                                                                      
###Don't change this comment - YaST2 identifier: Original name: xen###                                                
title Xen -- openSUSE 11.3 - 2.6.34-12                                                                                
    root (hd0,4)                                                                                                      
    kernel /boot/xen.gz vgamode=0x31a 
    module /boot/vmlinuz-2.6.34-12-xen root=/dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part5 resume=/dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part7 splash=silent quiet showopts vga=0x31a
    module /boot/initrd-2.6.34-12-xen

###Don't change this comment - YaST2 identifier: Original name: windows 1###
title windows 1
    rootnoverify (hd1,1)
    makeactive
    chainloader +1

###Don't change this comment - YaST2 identifier: Original name: windows 2###
title windows 2
    map (hd1) (hd0)
    map (hd0) (hd1)
    rootnoverify (hd1,0)
    makeactive
    chainloader +1

###Don't change this comment - YaST2 identifier: Original name: Linux other###
title Linux other
    rootnoverify (hd2,0)
    chainloader +1

If you mean force mounting an ntfs partition that linux is refusing to mount - no, don’t do it. It
usually means that the partition was not properly umounted (in windows) and the log still contains
unfinished transactions that have to be replayed.

That’s good to know. Knoppix couldn’t read it anyways so…

I miss the “extended partition table” text in yours. Maybe it is alright, as fdisk sees all
partitions. It does, no?

Yes, it shows all the partitions that I know of that exist.

In your log I see, when it is installing grub

Root device: /dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part5 (/dev/sdb5) (mounted on / as ext4)

Resume device: /dev/disk/by-id/ata-Maxtor_6Y120P0_Y41EF6NE-part7 (/dev/sdb7)

Root on sdb5? There is no such thing, sdb has only two partitions; it is writing on what fdisk says
it is sda5. If it then writes grub to sda1 it will be destroyed!

Yeah, it’s a bit confusing. That was from before sda and sdb swapped. So that’s actually correct because that’s the current sda. (as viewed from openSUSE – Knoppix still shows it as sdb)

user · July 30, 2010, 1:14am

On 2010-07-29 23:06, Caglow wrote:
>
> In any case, Knoppix couldn’t “see” the partition either. I did “fix”
> the Windows not loading issue. Apparently, sometime during all this
> “chaos”, the boot flag for sda had been set to sda1 instead of the
> Windows partition which was sda2. Oh, and the weird thing is that
> Knoppix sees the “correct” sda and sdb drives like openSUSE saw before
> the upgrade.

Not only that. It is flagged bootable, and there is a grub in there:

/dev/sda1: x86 boot sector; GRand Unified Bootloader, stage1 version 0x3, LBA flag 0x1, 1st sector
stage2 0xb54d2a5, GRUB version 0.97, code offset 0x48, OEM-ID "NTFS ", sectors/cluster 8,
reserved sectors 0, Media descriptor 0xf8, heads 255, hidden sectors 63, dos < 4.0 BootSector (0x80)

And it is NTFS.

When grub is installed in a partition, it writes a bit of code in the first sector, yes, (stage 1)
but also beyond that, I think the stage 1.5, and I don’t know where goes stage 2. The problem is
that some filesystems have structures there. so the operation destroys them (xfs is reported
problematic).

Look:

Elessar:~ # l /boot/grub/*stage1_5
-rw-r–r-- 1 root root 10084 Oct 24 2009 /boot/grub/e2fs_stage1_5
-rw-r–r-- 1 root root 9248 Oct 24 2009 /boot/grub/fat_stage1_5
-rw-r–r-- 1 root root 8556 Oct 24 2009 /boot/grub/ffs_stage1_5
-rw-r–r-- 1 root root 8552 Oct 24 2009 /boot/grub/iso9660_stage1_5
-rw-r–r-- 1 root root 10140 Oct 24 2009 /boot/grub/jfs_stage1_5
-rw-r–r-- 1 root root 8700 Oct 24 2009 /boot/grub/minix_stage1_5
-rw-r–r-- 1 root root 11336 Oct 24 2009 /boot/grub/reiserfs_stage1_5
-rw-r–r-- 1 root root 8828 Oct 24 2009 /boot/grub/ufs2_stage1_5
-rw-r–r-- 1 root root 8136 Oct 24 2009 /boot/grub/vstafs_stage1_5
-rw-r–r-- 1 root root 11072 Oct 24 2009 /boot/grub/xfs_stage1_5

Do you see ntfs there? No? Me neither… and that’s a bad thing. Grub can not be installed on an
NTFS partition… but you have.

I don’t like this. No, I don’t

–
Cheers / Saludos,

Carlos E. R.
(from 11.2 x86_64 “Emerald” GM (Elessar))