Hi
Vadim Vygonets wrote:
>
> Quoth Vadim Smelyansky on Wed, Jun 09, 1999:
> > On Sat, 22 May 1999, Isaac Aaron wrote:
> >
> > IA> "Lack of a Journaling file system -
> > IA> file system may not recover after
> > IA> unplanned downtime"
> > IA>- Since I don't know of any journaling capabilities
> > IA>of the ext2, I suppose they're right.
> > What rigth? second statement - lie. I tried several times to press
> > reset - its recover - beleave me, and NT does not :)
>
> Not always. Sometimes, when you press reset, Linux filesystem
> ends up in a completely broken state. It doesn't happen much,
> though -- happened to me only several times.
I must add, that there are things that e2fsck can't fix at all - e.g.
I once had a hardware problem, and fsck moved some files to
/lost+found, with permissions that didn't let me delete them
no matter what I did. I had to remove them manually with debugfs.
OTOH, just today my colleague reinstalled NT on a server with 4
data disks, and the new NT could not recognize 2 of them (each
had an NTFS partition on the whole disk). You can't fsck (chkdsk?)
an unrecognized fs, can you?
>
> > Usually I silently press reset on linux box to answer customer
> > question: "What happened if power go away?"
>
> Hmm. Usually I just silently cut my hand with a knife to answer
> customer question: "What happened if you are not careful and cut
> yourself accidentially?". Surely, the body will heal itself in
> few days, but it's not the desirable situation.
Although I agree, if you do it after 30 (as a default) seconds of
total inactivity, and let update sync, you are (almost) completely
safe.
>
> Vadik.
>
> --
> Any language that involves exposing private parts to friends is a
> tad suspect...
> -- Geoff Lane in the Monastery, about C++
>
>
didi