lkml.org 
[lkml]   [1998]   [Jun]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: e2fs + large disk == slug?
From
   Date: 	Tue, 16 Jun 1998 10:45:46 +0000
From: Richard Jones <rjones@orchestream.com>
Rogier Wolff wrote:
> Mark Lehrer wrote:
> > I am not a filesystem expert, but I would like to understand what
> > makes e2fs slow on my 4.5 gig partition; and if someone is working on
> > performance and would like some help, please holler. Some slow
> > operations, kernel 2.0.34:
> >
> > 1) removing 500 megs of files took 30 seconds.
> > 2) fsck of course.
>
> 3) Mounting large filesystems takes a long time: 46 seconds / 12G
The fsck is a real problem. I can see it taking a long time
when we start to fill the disk. Apart from writing a journalling
filesystem from scratch, has anyone got any suggestions about
how to improve fsck times?

Well, here are the answers.

1) I presented a paper at the Linux Expo and the Linux Kongress which
discussed the design of how to add B-tree support to the ext2
filesystem. Adding B-tree support to directories and for block
extent maps will significantly improve a large number of filesystem
operations, including large-scale deletes.

2) Stephen Tweedie presented a paper in the Linux Expo about adding
a metadata log to the ext2 filesystem. He has implemented a
prototype already, and will be integrating this functionality in the
2.3 kernel series.

3) To shorten the amount of time to mount a filesystem, you can use the
mount option "check=none". This will stop the filesystem from doing
a sanity check over the inode and block bitmaps, which can take
quite a bit of time on large filesystems.

- Ted

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu

\
 
 \ /
  Last update: 2005-03-22 13:43    [from the cache]
©2003-2011 Jasper Spaans