[gentoo-user] memory(gray matter) needs jog

[gentoo-user] memory(gray matter) needs jog

[Maxim Wexler]

Hi group,
A while back I needed help with an ext2 file system that required
checking every boot before mounting. The drive suffered from errors
involving 'non-contiguous files'.The solution was to run (this is
where things get hazy) e2fsck <option> <option> on the offender. I
keep thinking it's '-i -o', but there's no such options. Maybe it
wasn't e2fsck at all. I can't remember. I don't think it was
'-f''(force).

Anybody know what I'm talking about?

Maxim

Re: [gentoo-user] memory(gray matter) needs jog

[Albert Hopkins]

On Thu, 2009-10-29 at 15:54 -0600, Maxim Wexler wrote:
> Hi group,
> A while back I needed help with an ext2 file system that required
> checking every boot before mounting. The drive suffered from errors
> involving 'non-contiguous files'

I'm not sure what your problem was but this wasn't it.  "Non-contiguous
files" simply means there is some fragmentation in your filesystem,
which is totally natural.  It's not an error, or even a warning for that
matter. It's informational.  High fragmentation is an indication that
your filesystem may not be performing optimally, but it's not an error.

> .The solution was to run (this is
> where things get hazy) e2fsck <option> <option> on the offender. I
> keep thinking it's '-i -o', but there's no such options. Maybe it
> wasn't e2fsck at all. I can't remember. I don't think it was
> '-f''(force).

I'm not sure why your system needed to be checked for each boot.
Perhaps you can post the exact error message?  I'm pretty sure it wasn't
fragmentation.  What it *might* be saying (but again we can't verify
without an error message) is that your filesystem contains errors that
cannot be fixed in non-interactive mode (i.e. you need to run fsck
manually).

It's hard to tell you what option you need when we are unsure what
problem you are trying to fix.

[gentoo-user] Re: memory(gray matter) needs jog

[walt]

On 10/29/2009 03:46 PM, Albert Hopkins wrote:
> On Thu, 2009-10-29 at 15:54 -0600, Maxim Wexler wrote:
>> Hi group,
>> A while back I needed help with an ext2 file system that required
>> checking every boot before mounting...

> I'm not sure why your system needed to be checked for each boot...

Boy, my 'little gray cells' need a tonic, too.  I've been using ext3
(i.e.with journaling) for so long I can't even remember using ext2.

Wasn't it normal in the old days to fsck an ext2 fs with every boot?

Re: [gentoo-user] Re: memory(gray matter) needs jog

[Albert Hopkins]

On Thu, 2009-10-29 at 17:05 -0700, walt wrote:
> > I'm not sure why your system needed to be checked for each boot...
> 
> Boy, my 'little gray cells' need a tonic, too.  I've been using ext3
> (i.e.with journaling) for so long I can't even remember using ext2.
> 
> Wasn't it normal in the old days to fsck an ext2 fs with every boot?
> 

You should put your drink down.  And maybe go for a walk :)

No, I'm not aware of it *ever* being standard to fsck on every boot.
Perhaps you were using a bad distribution?  True, Linux does run fsck on
every boot, but fsck exits immediately if it determines your filesystem
was unmounted cleanly on the last shutdown.  It's only if you force a
fsck (e.g. with -f or /forcefsck) that it will run fsck on a clean
filesystem.

And remember, ext2 has *no* journal, so fsck was always very very slow.
To run it every time on a reboot would have been so painful that I
believe ext3 would have been invented in the early '90s instead of
2001 :)

Re: [gentoo-user] Re: memory(gray matter) needs jog

[Albert Hopkins]

On Thu, 2009-10-29 at 20:21 -0400, Albert Hopkins wrote:
> It's only if you force a
> fsck (e.g. with -f or /forcefsck) that it will run fsck on a clean
> filesystem. 

I stand corrected, it also forces a fsck after the user-tunable maximum
time between fscks.

-a