JFS on 2.2.x (was RE: [plug] Long kernel-linked PCMCIA delay?)

[Andrew Furey]

> I thought ext3 support in 2.2.x kernels stopped at
> 0.0.3, and they're up
> to something like 1.0.4 now.
> Can anyone shed further light on this as I'd prefer
> to use ext3, but I'm
> limited to 2.2 kernels for PPTP masquerading amongst
> other things.

I haven't been following it lately (other than to
automatically use ext3 on any 2.4 kernels I install)
but the patch I use can be gotten here:

ftp://terminus.net.au/unix/linux/ext3-0.0.5c.tar.gz

I use the 0.0.5c version although this tarball also
includes 0.0.5d, which I probably should be using :)


> paraphasing - "SOFTWARE RAID1 and RAID5 WILL NOT
> WORK" (their caps)
> & HW Raid should work as per normal (ext2)...

Correct, at least for RAID1. Basically the RAID
rebuilding process triggers ext3's builtin debugging
and causes kernel panics. Badness... this is fixed in
2.4 though.


> Default Journal for a while was 8MB for RAM reasons
> but they've upped it
> to 15MB -- not much cop for a v.busy 60GB+ drive
> IMHO.

Don't know about this - I create the journal using dd
(see the README in the tarball), I've had up to about
30Mb with no problems.

> All ext3 kernals ATM are debug with kbd, meaning
> "PAUSE" key will go to
> the debug mode and from there, you're in the bowels
> of the kernal as far
> as tracing goes (big hint... "g" resumes, just like
> the Mac programmers
> screen)

Note that for the above, you don't _need_ to apply the
kdb patches - just skip those parts (apart from the
Makefile, they all have "kdb" somewhere in the
filename). You don't really need it, I've found ext3
to be perfectly stable as it is.

Hope this helps.

Andrew


=====
"One Architecture, One OS" also translates as "One Egg, One Basket".

http://greetings.yahoo.com.au - Yahoo! Greetings
- Send your Valentines love online.