[GRLUG] WMNTUG Windows 7 Meeting

[john-thomas richards]

On Wed, Sep 16, 2009 at 03:02:59PM -0400, Adam Tauno Williams wrote:
> On Wed, 2009-09-16 at 14:35 -0400, john-thomas richards wrote:
> > On Wed, Sep 16, 2009 at 02:26:03PM -0400, Adam Tauno Williams wrote:
> > > On Wed, 2009-09-16 at 13:55 -0400, john-thomas richards wrote:
> > > > On Wed, Sep 16, 2009 at 01:46:40PM -0400, Adam Tauno Williams wrote:
> > > > > > on boot when ext3 hit its 35-mount automatic fsck. (On a rather full
> > > > > > 1TB drive, that can take long enough so as to look like it's not
> > > > > > getting anywhere...) 
> > > > > Ugh, I know!  You can change that with tune2fs.
> > > > How necessary is this anymore?
> > > To delay, or control, the mandatory filesystem check that occurs every X
> > > number of mounts.  On large filesystems this can take a very long time
> > > and on laptops it can just be down right annoying.
> > I know what it is, and certainly know how annoying it can be.  (Why is
> > it always the Xth number of mounts when you need to boot quickly?)  My
> > question isn't about tune2fs.  Is an automatic fsck necessary with a
> > journaling filesystem like ext3?
> 
> Define "necessary".  It is the only way to determine if something
> otherwise undetected has gone wrong - thus for the filesystem to be
> "trusted" it is "necessary",  if you have "faith" in the journal system
> [and hardware!] to correct/catch all issues then it is not "necessary".

Is it necessary in that one who actually has a clue about how such
things (filesystems) work and who understands system maintenance and
best practices far better than I would actually be compelled to have an
automatic fsck after so many boots?  Once again I am probably revealing
my profound ignorance of the deeper internals of Linux but I thought a
journaling filesystem went a long way to prevent filesystem corruption.
Thus my question if an automatic fsck is necessary.  I really don't
know.  You and many others on this list dwarf my knowledge in these
areas so I threw out the question.

> Of course we are only dealing with the "trust" in the filesystem
> meta-data here.  Integrity beyond that is a whole other kind of thing.

I don't have the chops to comprehend what this really means.  I get
(sort of) the concept of the meta-data, but beyond that...nope.

Hah!  I just noticed my random .sig is so very relevant to me and this
conversation!  Oh, the irony.
-- 
john-thomas
------
There's no sense in being precise when you don't even know what you're
talking about.
John von Neumann