Journaled File System

On Thu, 2010-10-14 at 10:54 -0400, Bhaskar, K.S wrote:
> I searched the archives, and found only one meaningful thread.  In that 
> case, the user of the file system planned to reformat the partition.  In 
> my case, I want to recover the partition and continue using it.
> 
> The jfs partition in question is the root partition of a Linux machine 
> that took a hard power down.  When I booted that partition, networking 
> didn't work, and I determined the problem to be a State NFS file handle 
> error with /etc/resolv.conf.  I was unable to delete the file, and 
> without /etc/resolv.conf, Linux (64-bit Ubuntu 10.04.1 LTS current with 
> patches) couldn't start networking.  The information below is with the 
> machine booted from an alternate partition, also 64-bit Ubuntu 10.04.1 
> LTS with jfsutils version 1.1.12-2.1 and the standard Linux kernel 
> 2.6.32-25 from the distro.
> 
> Thanks in advance for any and all help.

-ESTALE probably isn't the best choice for an error code, since the
problem has nothing to do with NFS.

Anyway, I would suggest backing up the filesystem, just to be safe, then
running "fsck -t jfs -f /dev/sda4".  Without the -f, fsck sees that the
filesystem is marked clean and doesn't actually look for or correct any
errors.

Shaggy
-- 
Dave Kleikamp
IBM Linux Technology Center

Thank you very much.  With -f, fsck worked like a charm.  (Confession: I 
didn't back the file system up first.  Since jfs has been rock solid for 
me for years and is the only file system I ever use for data that is 
important to me, I thought I wasn't living too dangerously!)

Regards
-- Bhaskar

GT.M - Rock solid. Lightning fast. Secure. No compromises.

On 10/14/2010 11:39 AM, Dave Kleikamp wrote:
> On Thu, 2010-10-14 at 10:54 -0400, Bhaskar, K.S wrote:
>  > I searched the archives, and found only one meaningful thread. In that
>  > case, the user of the file system planned to reformat the partition. In
>  > my case, I want to recover the partition and continue using it.
>  >
>  > The jfs partition in question is the root partition of a Linux machine
>  > that took a hard power down. When I booted that partition, networking
>  > didn't work, and I determined the problem to be a State NFS file handle
>  > error with /etc/resolv.conf. I was unable to delete the file, and
>  > without /etc/resolv.conf, Linux (64-bit Ubuntu 10.04.1 LTS current with
>  > patches) couldn't start networking. The information below is with the
>  > machine booted from an alternate partition, also 64-bit Ubuntu 10.04.1
>  > LTS with jfsutils version 1.1.12-2.1 and the standard Linux kernel
>  > 2.6.32-25 from the distro.
>  >
>  > Thanks in advance for any and all help.
>
> -ESTALE probably isn't the best choice for an error code, since the
> problem has nothing to do with NFS.
>
> Anyway, I would suggest backing up the filesystem, just to be safe, then
> running "fsck -t jfs -f /dev/sda4". Without the -f, fsck sees that the
> filesystem is marked clean and doesn't actually look for or correct any
> errors.
>
> Shaggy
> --
> Dave Kleikamp
> IBM Linux Technology Center
>

_____________

The information contained in this message is proprietary and/or confidential. If you are not the intended recipient, please: (i) delete the message and all copies; (ii) do not disclose, distribute or use the message in any manner; and (iii) notify the sender immediately. In addition, please be aware that any message addressed to our domain is subject to archiving and review by persons other than the intended recipient. Thank you.
_____________

On Thu, 2010-10-14 at 11:49 -0400, Bhaskar, K.S wrote:
> Thank you very much.  With -f, fsck worked like a charm.  (Confession: I 
> didn't back the file system up first.  Since jfs has been rock solid for 
> me for years and is the only file system I ever use for data that is 
> important to me, I thought I wasn't living too dangerously!)

In the future I'd recommend at least backing up the directory containing
a known problem (in this case /etc).  When fsck does need to fix
something, it has a tendency to throw away the directory and send all
the files to /lost+found/.  In this case the directory was sane, but the
inode containing the file must have been the problem.
-- 
Dave Kleikamp
IBM Linux Technology Center

GT.M - Rock solid. Lightning fast. Secure. No compromises.

On 10/14/2010 11:59 AM, Dave Kleikamp wrote:
> On Thu, 2010-10-14 at 11:49 -0400, Bhaskar, K.S wrote:
>  > Thank you very much. With -f, fsck worked like a charm. (Confession: I
>  > didn't back the file system up first. Since jfs has been rock solid for
>  > me for years and is the only file system I ever use for data that is
>  > important to me, I thought I wasn't living too dangerously!)
>
> In the future I'd recommend at least backing up the directory containing
> a known problem (in this case /etc). When fsck does need to fix
> something, it has a tendency to throw away the directory and send all
> the files to /lost+found/. In this case the directory was sane, but the
> inode containing the file must have been the problem.

[KSB] Fair enough!  Actually, I do have some redundancy (paranoia) in my 
process.  When I set up a computer, I always set it up with two alternate 
root partitions that I can boot into.  If anything goes wrong with one 
(e.g., an upgrade gone awry), I can always use rsync (followed by a 
little editing, e.g., of /etc/fstab) from the other root to recover a 
partition.  A spare root may be a few GB of overhead, but it does come in 
handy from time to time.

Thanks again for your help, and for an excellent filesystem.

Regards
-- Bhaskar

_____________

The information contained in this message is proprietary and/or confidential. If you are not the intended recipient, please: (i) delete the message and all copies; (ii) do not disclose, distribute or use the message in any manner; and (iii) notify the sender immediately. In addition, please be aware that any message addressed to our domain is subject to archiving and review by persons other than the intended recipient. Thank you.
_____________