[KLUG Members] Defrag'ing ext3 filesystems?

[Peter Buxton]

On Thu, Oct 03, 2002 at 09:00:27AM -0400, Adam Williams wrote:

> Thats great!  I must have missed that one, as I've been praying for
> feature freeze to start soon.  Of course, 2.5.x has undergone
> meta-evolution since 2.4.x, so the feature freeze may last awhile.  And
> we all remember 2.4.x prior to 2.4.5 (eeeeEEE, YIKES!).

I would note that the big changes added to the kernel this go round have
been made by people who don't seem to have personality problems. Ingo
Molnar isn't Rik van Riel (RvR isn't anymore, either, and his latest
kernel patches have been accepted), Andrew Morton cheerfully folded his
low-latency patches into Robert Love's kernel preemption, ALSA, UML and
XFS all merged without a hiccup.

> Right I wasn't arguing that it didn't work.  Just that it isn't a "real"
> solution relative to the most common current uses for Linux in the
> enterprise.  ext2/3 is gradually becoming Linux's dirty little secret.

Cool by me. I'd rather have an indestructible FS that needs maintenance
for performance than an invasive FS that causes kernel panics. It's
interesting to watch foreign-coded projects get Linuxized: van Riel
rewrote FreeBSD's VM for 2.4 Linux, and it got replaced by Andrea
Arcangeli's VM, which was a 2.2-ish VM with the 2.2 limitations removed
to allow Linux's new features (SMP, &c.) to work. 

Speaking of SGI:

http://www.osnews.com/story.php?news_id=1859

See what another company, famous for graphics work, did with their X
server.

-- 
http://www.killdevil.org/~peter
Power tools for power fools.