Hello,

thanks for looking into this.

On Apr 7, 2008, at 13:40, Colin King wrote:
> 1. Did the running of rdiff-backup trigger the reported Oops and  
> cause the broken filesystem?
> 2. Or was the oops caused later when you tried to do rdiff-backup  
> on an already broken filesytem?

The oops was caused by running rdiff-backup. I'll give an overview of  
how this happened:

I had been using the disk occasionally for backups using rdiff-backup  
under OS X (Macbook). Since I needed to free space on the partition,  
I tried deleting old increments with rdiff-backup from OS X, but  
aborted this because it took too long. It is possible that I also  
aborted the last backup run before this. Since I suspected the  
Macbook to be at fault for the slowness, I decided to delete the  
increments from my Ubuntu desktop. I disabled journaling from OS X  
before, to allow rewrite mounting. While running "rdiff-backup -- 
remove-older-than", the oops occurred.

I believe the filesystem was still ok at that point, though I can't  
be absolutely sure. I think I "verified" it using OS X's disk utility  
shortly before (that's an fsck, I guess).

> A few more questions:
>
> Is the filesystem salvageable with fsck?

I ran fsck on OS X just now, and it aborts with

BAD SUPER BLOCK: MAGIC NUMBER WRONG

Should I run a different fsck?

> Can you re-trigger this kernel Oops message?

I haven't tried so far. Should I keep the broken file system around,  
or can I use that disk for testing?

> I've tried to reproduce this bug, but as yet, not succeed. Any
> information on the kind of structure of the tree may be helpful (e.g.
> depth of directory tree, etc..) to try and help me reproduce this.

About all I can say is that it was a backup of an OS X installation.  
I don't know at which point the error occurred.

...

There's also the possibility that there's something wrong with the  
USB disk, though it has appeared to work fine at least attached to  
the desktop and another laptop. With the Macbook, I had the problem  
that after a while of running rdiff-backup, IO performance got hugely  
worse (for the whole system, even after detaching the USB drive).  
This is the rdiff-backup-slowness I mentioned above. I tend to think  
this is rather the fault of the Macbook, or its internal hard disk,  
which has since failed. I can't say whether the slow-down still  
occurs with the new drive yet.

Cheers
Rob

-- 
hfsplus oops, corrupted file system
https://bugs.launchpad.net/bugs/202595
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.

-- 
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to