Grub error 17 & reinstalling GRUB

Hi. You can read my previous post here:

Problem with partition table - openSUSE Forums

It contains my partition table. I made only a slight change today - I took 10GB from my biggest NTFS partition and assigned it to opensuse’s root partition (because it was too small). I used Paragon Partition Manager 9 to do it (in Windows XP)

But when I rebooted, I got “GRUB Error 17” message.

I ran opensuse’s repair dvd and it told me there’s file system’s incosistency (or something like that) in both root and home partitions. I clicked to repair it, but it repsponded with:


resize inode not valid run fsck manually

For both partitions.

You can read from my previous topic that I’ve got one 20 GB primary partition. I wanted to overwrite MBR with Windows’ code, and then set this partition as the active one (I know, I’m a noob, but I hoped I could manage that)

I ran Windows XP Repair Console and wrote “fixmbr” but after rebooting and checking for CD, PC did nothing. Then I logged once again, logged into my 20 GB drive, and wrote “fixboot”, but it said it can’t recognize partition or something like that.

So I am stuck with nothing, literally - what I need to do is to reinstall GRUB, make the 20 GB partition active, and then repair the filesystems on linux partitions.

I know I have really bad luck with partitions (second serious problem in a quite short time) - just tell me, how in the future I’m supposed to change the linux partitions’ sizes.

All help is greatly appreciated!

You should be able to fix the system using the suse install disk. Choose the repair option on the menu.
To run fsck manually:
Again use the install disk.
Choose the rescue option to get a console.
Type fdisk -l to get your partition info.
Run fsck -p /dev/drive_info where drive info is the partition you want to check.

Alternatively, you can use a live distro cd to do the same.

Ok, I boot into livecd, and I need to check sda8 and sda9


fsck -p /dev/sda8

/dev/sda8: Resize inode not valid

/dev/sda8: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
(i.e. without -a or -p options)

The same with sda9.


fsck /dev/sda8

e2fsck 1.41.1
fsck.ext3: Group descriptors look bad... trying backup blocks...

Resize inode not valid. Recreate<y>? yes

/dev/sda8 was not cleanly unmounted, check forced.
Pass1: Checking inodes, blocks and sizes
Group 0's inode table at 4 conflicts with some other fs block. 
Relocate<y>?

Group 0's inode table at 5 conflicts with some other fs block. 
Relocate<y>?

Group 0's inode table at 6 conflicts with some other fs block. 
Relocate<y>?

And I clicked “y” until 14. Then I Ctrl+C the process. Or should I have continued it?

Of course, the same happens to sda9.

PS.: Ok, I’m trying to fix sda8 manually. I’m already holding “y” like for 10 minutes. Wish me luck !

Have a look at the man pages. Execute this command in a console: man fsck. Look particularly at the options: -a and -y. Maybe one will allow you not to hold the “y” key for 10 minutes. lol!

I thought about it, but each time I use any option, it says:


/dev/sda8: Resize inode not valid

/dev/sda8: UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY.
(i.e. without -a or -p options)

I’ve put some weight on the “y” button, and I let it go like this for about 4 hours.

I now know why the fsck command is pronounced “f-suck” by some people. It takes awfully long time to repair the partition. It went through ~600k inodes, and I had no idea how much more there could be. “df -i” should say, but it says something is corrupted and it can’t tell. And it’s only 15 GB partition! Windows checked total of 250GB volumes during 2 minutes (after I ran it after resizing)!

Because I have no idea how long the process could take, and because I’m afraid to let the computer click “yyyyyyyy” for the whole night (it might agree to some disk-harming option) I think the best option would be to give in and just reinstall opensuse.

Just tell me, how am I supposed to resize linux partitions in the future if I still want it to work afterwards?