Installation must've split my home partition

[cxd]

Hi all, I'm having a rather peculiar problem. I had a reiserfs home partition and I wanted to install the 11.1 openSuse distro over. The partition is about 140 GB in size and all worked very well, until it came to assigning a mount point to the partitions when I noticed that for some strange reason instead of /dev/sda6 I had two partitions (/dev/sda6 and /dev/sda7) with 20 GB and 120 GB each. Running fsck I actually found out that /dev/sda7 wasn't a real partition, while /dev/sda6 had errors (block count was off). So I thought well, just rewrite the partition table as I remember it to be and all should be well. But that was apparently a wrong conclusion, as it still complained about the super block being erroneus. reiserfsck now aborts after some badblocks (which I never noticed before) and I still can't mount the partition. So what I'm doing right now is I've run

badblocks -n -b 4096 -o badblocksfile /dev/sda6

and then

reiserfsck --fix-fixable --badblocks badblocksfile /dev/sda6

But I've no good feeling at all about it.

Is there a way to temporarily mount the partition so I can get some data out of it?

[caf4926]

A live cd should let you mount the partition. A suse one will do, mind you I always use knoppix or mepis.

[FeatherMonkey]

I wouldn't mount it until you're sure the disk isn't buggered. I would look at ddrescue create an image and try to mount that. Though perhaps some off the blind commands have made some of the data un-retrievable.

Though it is feasible to mount ro but many live distro's will mount or probe the disk or try to use swap etc..

What this means is anyone's guess "just rewrite the partition table as I remember it to be" in theory had you changed the structure of the table you can change it back and it will be OK.

But...
You had not made any changes you where still at the choosing stage, so either you're not telling us something or the disk is dying.

So if this was me(I would of one had a backup of data I cared for! Had I got to this stage you begin looking at data retrieval. So the first thing is to get a dump and try all fixes on that.)

Will you be able to reverse it who knows but I suspect you have a dying disc. Why, well I presume and don't really care enough to confirm. That the partition table is stored in the early bit of the disk(Maybe mbr(Note not sda6/7)), you seem to be playing with badblocks. And each time you mount it, probe it, try to write to it, and in certain cases even reading it, you give it the opportunity to lose more data.

[Knurpht]

I agree completely with Feathermonkey. Too much has already been done, to be able to say something decent about it. Either you already messed up, or the disk is at it's end.
Put the disk in another linux-box together with a backupdisk and use dd or ddrescue to retrieve (some of) your data, but from what I've read something like data-integrity will not apply.

[cxd]

I actually got it working again: Brian's Blog: Reiserfs - Dealing with Bad Blocks
I guess I was lucky ^^