From: "Rechenberg, Andrew" <ARechenberg@shermanfinancialgroup.com>
To: "'linux-mm@kvack.org'" <linux-mm@kvack.org>
Subject: kupdated high load with heavy disk I/O
Date: Wed, 14 Nov 2001 18:01:23 -0500 [thread overview]
Message-ID: <35F52ABC3317D511A55300D0B73EB8056FCC19@cinshrexc01.shermfin.com> (raw)
Hello,
I have read some previous threads about kupdated consuming 99% of CPU under
intense disk I/O in kernel 2.4.x on the archives of linux-kernel (April
2001), and some issues about I/O problems on linux-mm, but have yet to find
any suggestions or fixes. I am currently experiencing the same issue and
was wondering if anyone has any thoughts or suggestions on the issue. I am
not subscribed to the list so would you please CC: me directly on any
responses? I can also check out the archives at theaimsgroup.com if a CC:
would not be appropriate. Thank you.
The issue that I am having is that when there is a heavy amount a disk I/O,
the box becomes slightly unresponsive and kupdated is using 99.9% in 'top.'
Sometimes the box appears to totally lock up. If one waits several seconds
to a couple of minutes the system appears to 'unlock' and runs sluggishly
for a while. This cycle will repeat itself until the I/O subsides. The
memory usage goes up to the full capacity of the box and then about 10MB of
swap is used while this problem is occurring. Memory and swap does not get
relinquished afer the incident.
The issue appears in kernel 2.4.14 compiled directly from source from
kernel.org with no patches. These problems manifest themselves with only
one user doing heavy disk I/O. The normal user load on the box can run
between 350-450 users so this behavior would be unacceptable because the
application that is being run is interactive. With 450 users, and the same
process running on a 2.2.20 kernel the performance of the box is great, with
only a very slightly noticeable slow down.
I am running the Informix database UniVerse version 9.6.2.4 on a 4 processor
700MHz Xeon Dell PowerEdge 6400. The disk subsystem is controlled by a PERC
2/DC RAID card with 128MB on-board cache (megaraid driver compiled directly
in to the kernel). Data array is on 5 36GB 10K Ultra160 disks in a RAID5
configuration. The box has 4GB RAM, but is only using 2GB due to the move
back to the 2.2 kernel. The only kernel paramters that have been modified
are in /proc/sys/kernel/sem. All filesystems are ext2.
If you need any more detailed info, please let me know. Any help on this
problem would be immensely appreciated. Thank you in advance.
Regards,
Andrew Rechenberg
Network Team, Sherman Financial Group
arechenberg@shermanfinancialgroup.com
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/
next reply other threads:[~2001-11-14 23:01 UTC|newest]
Thread overview: 8+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-11-14 23:01 Rechenberg, Andrew [this message]
2001-11-15 4:30 ` John McCutchan
2001-11-19 16:23 ` Marcelo Tosatti
2001-11-20 22:58 Rechenberg, Andrew
2001-11-26 15:08 Rechenberg, Andrew
2001-11-26 15:29 ` Andrea Arcangeli
2001-11-26 23:07 ` Ken Brownfield
2001-11-30 15:13 Rechenberg, Andrew
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=35F52ABC3317D511A55300D0B73EB8056FCC19@cinshrexc01.shermfin.com \
--to=arechenberg@shermanfinancialgroup.com \
--cc=linux-mm@kvack.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