Bootup Superblock Issue Related to /dev/disk/....-part8

Some time ago, I decided I needed to expand my openSUSE partition because I was running out of room. There was some unallocated space following my /home and swap partitions that I wanted to assign to my / partition.

So after taking note that bad things could happen (as I’ve read everywhere), I got myself a copy of GParted (since Yast! partitioner doesn’t move partitions, does it?) to start getting the thing to work. Well, actually, I got PartedMagic which has GParted (I realize it could be a problem NOW if they didn’t have the latest version).

So the first step was to move the swap partition. I decided to do it one step at a time and only moved the swap partition to the end. That turned out fine…no problems whatsoever.

The next step is where the problem came in. It was moving the /home partition. It took a while to move but once it got to the end, it showed a message that was similar in nature to “error detected” without giving a long list of error messages. If it did, I probably would have copied it down. There wasn’t anything in the “logs” either in GParted besides what it showed. Looking back, I probably should’ve went into /vars/logs to get additional log information but I guess I wasn’t smart enough to do so (provided that GParted does leave log messages there which I think it does).

After that, it refreshed the drive and it ended up showing no partitions with an error and saying the only thing it can do is create a new partition table.

After that, openSUSE wouldn’t boot. After loading, (both normal and failsafe modes) it gives me the message in the link at the bottom. I can still access those partitions fine. Nothing’s corrupted. Windows also boots up fine (it’s on a different drive though) and reads the affected drive fine. Linux-based LiveCDs (including openSUSE 11.3) reads the partitions fine too.

I’ve tried using e2fsck on my ext4 partitions with commands I found during a search and they seemed to “fix” those partitions but it still won’t boot and gives me the same message. Looking at it carefully, it seems the reason it can’t find a superblock is because it can’t find part8 of whatever that thing under /dev/disk/by-id/ is.

I would very much prefer a failsafe (or at least mostly failsafe) solution that won’t (or is unlikely) to result in requiring the restoration of a backup but I do understand that there is always the chance of something going wrong that will kill everything. Considering that I can still access the data on that drive, I don’t believe the data is corrupted. Maybe the drive itself (as in whatever signatures it may leave) but not the filesystems.

This is what I end up with trying to boot into openSUSE (sorry, I don’t know how to get the log when it doesn’t put it in logs directory: https://docs.google.com/leaf?id=0B3Axt7C5WCJeZDE1ZTU1YzUtZjEzNC00ZjNmLThiYjctMzliNDBkNDFmOWI4&sort=name&layout=list&num=50

Bump. Anyone have any ideas or suggestions? Hope I didn’t scare anyone with the big post. If so, you can probably ignore it all and just see the image.

I suspect something went wrong with the move and resize. This may or may not be recoverable. (thats why backups exist) There may be a utility on the Parted Magic CD that may help. But the name escapes me (fixdisk???). I have never used it myself.

The strange thing though is that all the data is intact. I would expect that the data would be corrupted or something but I can still grab the data off (well, I’m getting only the important stuff…) the partitions. However, if at all possible, I’m looking for some kind of alternative that can somehow avoid clearing out the drive and recreating the partitions. Because I have a backup…but it’s over the internet on another server…and downloading all that (“only” 80GB but considering a relatively slow connection…) would take forever (Yes, I am looking into a better way to backup. Like to DVDs (since all the “dynamic” stuff is in subversion repositories). But it’s unlikely I’ll be able to do that as of yet (likely not until September will I be able to get a place with DVD burners and stuff).

FWIW, is your problem that you can’t dual boot? Partitions are fine as far as you know, OpenSuse data is there as well as Windows?

Your PC automatically boots to Windows, ie, you never see the Grub boot menu offering you the choice of booting into OpenSuse or Windows.

If that is the case then you need to fix Grub. I’ve found the following links helpful:
Grub From the Ground Up **
GRUB Manual - Installation**

The second link shows commands (2) o write Grub to the partition or write to the Master Boot Record (MBR).

Please backup your data (probably a good idea to test restore using the backup too!) and proceed with caution.

No that’s not the problem. The problem is, when I boot into openSUSE, the boot stops halfway with the load screen turning into what the image shows: https://docs.google.com/leaf?id=0B3Axt7C5WCJeZDE1ZTU1YzUtZjEzNC00ZjNmLThiYjctMzliNDBkNDFmOWI4&sort=name&layout=list&num=50

Its a secure web link (https://) that wants a login to google docs. Is there a way to post for the public without a google docs login?

Oh does it really? Sorry. I changed the settings so that you can view it without being logged in. Or at least I can view it without being logged in. I’ll post it elsewhere if it still refuses to work. https://doc-10-5k-docs.googleusercontent.com/docs/secure/ha0ro937gcuc7l7deffksulhg5h7mbp1/l4cjhn2d5vn4kjo98cshamoov8r6m0t4/1279994400000/11200873799262235585/*/0B3Axt7C5WCJeZDE1ZTU1YzUtZjEzNC00ZjNmLThiYjctMzliNDBkNDFmOWI4

Thank you.
From the new link, it looks to me like you are trying to fschk.ext4 your partition 8 but partition 8 is not formatted ext4, it is ext2.

Before **you try to run, e2fsck -b 8193, **hopefully some of the other posters can confirm this or provide another explanation.

You say you merged a couple of partitions. Did partition 8 get merged with partition 7 and not exist any more? That would prevent the filesystem checking script from finding it. If so, perhaps all you need to do is comment out the line relating to partition 8 from /etc/fstab. A rescue CD can let you do mount your / partition to do this edit.

But before you do this, you might want to check the layout of your disk. fdisk -l /dev/sdb (I think it’s sdb) should show you what partitions you have.

You say you merged a couple of partitions. Did partition 8 get merged with partition 7 and not exist any more? That would prevent the filesystem checking script from finding it. If so, perhaps all you need to do is comment out the line relating to partition 8 from /etc/fstab. A rescue CD can let you do mount your / partition to do this edit.

I did not merge any partitions. Or at least I don’t think I did. Did the swap partition merge with partition 7? (I can still access partition 7) But the other thing is that GParted can’t “find” any partitions on /dev/sdb even though I can access them fine. The openSUSE partitioner (of the LiveCD) also gives me an error saying that there’s a problem with disk and I can view the partitions but I can’t edit anything. (and that’s exactly what happens when it loads) But the swap partition does seem to have disappeared which probably explains the not found problem. I will try what you said and see if it boots up properly although even if it does, I still wouldn’t be able to edit the partitions (meaning no swap).

Thank you.
From the new link, it looks to me like you are trying to fschk.ext4 your partition 8 but partition 8 is not formatted ext4, it is ext2.

Before you try to run, e2fsck -b 8193, hopefully some of the other posters can confirm this or provide another explanation.

When I put that in with the /dev/disk/by-id/…-part8 (written in full of course), I end up with a not found error saying pretty much the same thing that the image is saying right now. And I’m not sure why its using that long name instead of say /dev/sdb8 or something similar. Actually, I think /dev/sdb8 was the swap partition. /dev/sdb6 and /dev/sdb7 are ext4 partitions. /dev/sdb1 is an ntfs partition. /dev/sdb2 is a logical partition containing /dev/sdb[6-8]. I did “fix” some “errors” in my two ext4 partitions though that seemed to have changed nothing. I used this for that: http://linuxexpresso.wordpress.com/2010/03/31/repair-a-broken-ext4-superblock-in-ubuntu/

Not having a swap partition is not serious if you have enough RAM. In fact, if you have enough RAM, swap will hardly be touched. You could even use a file on a normal partition as a swap device your want a bit of insurance. So let it boot up and then comment out that line.

How old is your version of gparted? Remember that openSUSE went to 256 byte inodes a while back and this could prevent older versions of tools from working on the filesystem.

How old is your version of gparted? Remember that openSUSE went to 256 byte inodes a while back and this could prevent older versions of tools from working on the filesystem.

So far, that seems to be the most likely reason. My version of Parted Magic (4.11) came with GParted 0.5.2 (I think).

Also, I made a typo saying that /dev/sdb2 is a logical partition. It’s actually an extended partition. (as would make sense from the way I stated it)

Have no idea about long name either though its a Maxtor, I lean towards hd.

Well the error says if it’s a swap files its a no-good for fsck.

If you can read /etc/fstab from Windows or something, what is the mount point for /dev/sda8?

Actually, it looks like you can login as root, and

cat /etc/fstab

. Or you may need to mount your / (root) partition then display the /etc/fstab.

Don’t do anything to hurt your hd, os or data. Just display stuff.

I will have to reboot into the LiveCD to open /etc/fstab since it’s an ext4 partition. (and I have no more room to install additional software) I’m pretty sure it’s not mounted though I’ll check just to be sure.

Yay, problem 1 of 2 has been solved. Thank you for your help with getting openSUSE working again. :slight_smile:

Not having a swap partition is not serious if you have enough RAM. In fact, if you have enough RAM, swap will hardly be touched. You could even use a file on a normal partition as a swap device your want a bit of insurance. So let it boot up and then comment out that line.

That seems to have worked. Apparently, partition 8 wasn’t even the swap partition. It claimed to be a ext4 partition but I don’t think I had any other ext4 partitions.

If you can read /etc/fstab from Windows or something, what is the mount point for /dev/sda8?

It was at /mnt/tmp. Not sure how it got there.

So commenting it out worked though now, at startup, I get a “plasma has crashed” error (even though it’s still running fine) maybe because the tmp directory isn’t being created or something.

So part 2 of 2 is the question of what happened to the disk? I guess it’s not mandatory I fill in that extra 10GB of unallocated space but I would at least like to know what’s the problem in case I ever decide to fix it (after getting a local backup device :wink: ). The error message from Yast! Partitioner goes as follows:

The partioning on disk /dev/sdb is not readable by the partioning tool parted which is used to change the partitioning table.

You can use the partitions on disk /dev/sdb as they are. You can format them and assign mount points to them but you cannot add, edit, resize or remove partitions from that disk with this tool.

Any idea what may cause that problem?

Oh, I missed the note about using fdisk. Well, I didn’t “miss” it actually, I just forgot about it. I “remembered” it again after doing a few searches. The results of fdisk -l seem interesting. One of the partitions (my swap) was moved “out of bounds”. I got this:

Warning: ignoring extra data in partition table 7
Warning: ignoring extra data in partition table 7
Warning: ignoring extra data in partition table 7

Disk /dev/sdb: 122.9 GB, 122942324736 bytes
255 heads, 63 sectors/track, 14946 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x0698a55e

Device Boot Start End Blocks Id System
/dev/sdb1 * 1 11750 94373842 7 HPFS/NTFS
/dev/sdb2 11751 14946 25671870 f W95 Ext’d (LBA)
/dev/sdb5 11751 12794 8385898+ 83 Linux
/dev/sdb6 12795 13708 7341056 83 Linux
/dev/sdb7 ? 253600 344197 727723910+ c0 Unknown

Also, plasma no longer claims to have crashed at startup. Instead, Amorak does that now. Weird. I’ll just Kaffeine since it works.

Again, thanks for your help everybody.

That’s good news, glad you resolved it, you’re booting into openSUSE as you expected?

No idea why you have a /media/tmp but it got mounted in /etc/fstab.

RE: plasma error – new to me.
If you are dual booting (Win7/openSUSE) normally, do you have a /tmp ? Maybe that’s the source of the plasma error?
If no /tmp make one as root

 
mkdir /tmp
chmod 1777 /tmp 

I have a /tmp folder and it seems to be working. I had some problems with Firefox (blank preference windows, etc) but closing it down and restarting it seems to have fixed the issue. All I’m left with now is the Amorak issue and the issue with the partition 7. So essentially, everything is working just as well as before. And yes, swap doesn’t seem to really make a difference in the speed.