All of lore.kernel.org
 help / color / mirror / Atom feed
From: Shirley Shi <shirley@kasenna.com>
To: linux-kernel@vger.kernel.org
Subject: All filesystems hang under long periods of heavy load (read and write) on a filesystem
Date: Mon, 03 Nov 2003 15:46:22 -0800	[thread overview]
Message-ID: <3FA6E8CE.6040208@kasenna.com> (raw)

Can anyone know why all filesystems hang under periods of heavy load on 
one of the filesystem? Once the filesystems hang, any command related to 
the filesystem, like 'ls', 'cat',etc., will stick forever until re-power 
cycling the machine.

I kept running the following script to read and write the data on a same 
filesystem(ext2 or XFS) since we need do some tests for the storage. Is 
half day, onn the beginning, the system was running well. But after 
running the script for a long time, such a half day, one day or two 
days,  all filesystems would get hung, including the root filesystem 
although I didn't do any heavy load on it. The file(M.1) I used for 
reading and writing is about 2.5GB.


@ total = 115
while (1)
  @ cc = 2
  while ($cc <= $total)
     dd bs=512k if=/data/M.1 of=/data/M.$cc
    echo "copying $cc  of   $total..."
    @ cc = $cc + 1
  end
  rm -f  /data/M.*
end


I tried RH8.0 with kernel 2.4.18 and kernel 2.4.21 with XFS and patch 
rmap15j. I have the same issue running with the two kernels. Basically I 
have two filesytems configured. One for the root configured with ext3, 
and another is for the data configured with ext2 or XFS. With either 
ext2 or XFS, I have the same problem.

I also tried on different Dual CPU machines as follows, but saw the same 
problem.

- A dual-CPU machine with an on-board 320 SCSI controller(running 
AIC79XX.o driver) connecting a disk drive as the root system and a 
MegaRAID 320-4X controller connecting to several disk drives. I created 
two H/W logical RAID devices and built the data filesystem with the S/W 
RAID0(/dev/md0).
- A HP dual-CPU machine with the HP Smart Array 5i connecting to a disk 
drive as the root system and a 160 SCSI controller connecting with 13 
disk drives and configured as a S/W RAID0(/dev/md0) the data filesystem.
- A dual-CPU machine with an on-board 320 SCSI controller(with AIC79XX.o 
driver) connecting a disk drive as the root system and a FC controller 
connecting to a RAID storage.

Any comment would be appreciated.

Thanks,

Shirley



             reply	other threads:[~2003-11-03 23:46 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-11-03 23:46 Shirley Shi [this message]
2003-11-04  7:33 ` All filesystems hang under long periods of heavy load (read and write) on a filesystem bert hubert
2003-11-04  9:28 ` Bernd Petrovitsch
2003-11-04 15:52 ` Randy.Dunlap
  -- strict thread matches above, loose matches on Subject: below --
2003-12-23 18:01 Ged Haywood
2003-12-29 15:32 ` Ged Haywood
2003-12-29 19:32 ` Marcelo Tosatti
2003-12-31 17:26   ` Ged Haywood

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=3FA6E8CE.6040208@kasenna.com \
    --to=shirley@kasenna.com \
    --cc=linux-kernel@vger.kernel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.