kernelnewbies.kernelnewbies.org archive mirror
 help / color / mirror / Atom feed
From: arlie@worldash.org (Arlie Stephens)
To: kernelnewbies@lists.kernelnewbies.org
Subject: Work
Date: Fri, 25 Jul 2014 15:23:42 -0700	[thread overview]
Message-ID: <20140725222342.GA556@worldash.org> (raw)
In-Reply-To: <CAPDOMVj=mX9hwvGpYzs2M9iBx6O+4nrENDJU=cG-u7Cqx0dJ2g@mail.gmail.com>

On Jul 25 2014, Nick Krause wrote:
> > But if nothing like that is jumping out at you, maybe you should go look
> > around and see if there's something in userspace that *does* jump out at you.
> 
> I am interested in file systems and will be working on brtfs and ext4.
> Cheers Nick

If you want an annoying problem, explain and/or fix directory
performance on ext4. I've got a server where an ls of a directory took
5 seconds, according to "time", even though it only has 295 entries at
present.

It's a rather large directory, 4 times the size of a normal directory
in the opinion of ls -ld, and filefrag reports it has 7 extents. 

I'm also getting anecdotal reports of rename() taking unreasonable
amounts of time, probably with at least one of the directories
involved being in a similar state.  (They would have been created and
populated by the same software.) 

Probably the directory had a large number of files and/or
subdirectories some time in the past, and this is "expected
behaviour".  Expected or not, it seems to me that it's about time that 
linux handles directories with something closer to the efficiency with 
which it handles files ;-) 

-- 
Arlie

(Arlie Stephens					arlie at worldash.org)

  reply	other threads:[~2014-07-25 22:23 UTC|newest]

Thread overview: 30+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2014-07-24 16:38 Work Nick Krause
2014-07-24 16:43 ` Work Kristofer Hallin
2014-07-24 16:44 ` Work Lucas Tanure
2014-07-24 16:47   ` Work Nick Krause
2014-07-26 21:13   ` Work Yi Li
2014-07-26 21:55     ` Work Lucas Tanure
2014-07-24 16:51 ` Work Andev
2014-07-24 17:10   ` Work Nick Krause
2014-07-25  2:23     ` Work Nick Krause
2014-07-25  5:33       ` Work ravi ranjan Mishra
2014-07-25 11:44         ` Work Lucas Tanure
2014-07-25 12:17           ` Work Robert P. J. Day
2014-07-25 15:19             ` Work Nick Krause
2014-07-25 17:18               ` Work Robert P. J. Day
2014-07-25 17:28         ` Work Valdis.Kletnieks at vt.edu
2014-07-25 17:42       ` Work Valdis.Kletnieks at vt.edu
2014-07-25 21:54         ` Work Nick Krause
2014-07-25 22:23           ` Arlie Stephens [this message]
2014-07-25 23:02             ` Work Nick Krause
2014-07-25 23:35             ` Work Valdis.Kletnieks at vt.edu
2014-07-25 23:44               ` Work Nick Krause
2014-07-26  1:08               ` Work (really slow directory access on ext4) Arlie Stephens
2014-07-26  1:22                 ` Nick Krause
2014-07-30  2:34                   ` Nick Krause
2014-07-30 17:38                     ` Arlie Stephens
2014-07-30 19:48                       ` Valdis.Kletnieks at vt.edu
2014-07-30 20:45                         ` Nick Krause
2014-07-31 23:36                           ` Arlie Stephens
2014-07-31 23:41                             ` Henry Hallam
2014-08-01  1:47                               ` Nick Krause

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20140725222342.GA556@worldash.org \
    --to=arlie@worldash.org \
    --cc=kernelnewbies@lists.kernelnewbies.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).