public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Jesse Pollard <pollard@admin.navo.hpc.mil>
To: vda@port.imtp.ilyichevsk.odessa.ua,
	"Jim Sibley" <jlsibley@us.ibm.com>,
	linux-kernel@vger.kernel.org
Cc: riel@conectiva.com.br, ltc@linux.ibm.com,
	"Troy Reed" <tdreed@us.ibm.com>
Subject: Re: Killing/balancing processes when overcommited
Date: Fri, 13 Sep 2002 07:54:21 -0500	[thread overview]
Message-ID: <200209130754.21751.pollard@admin.navo.hpc.mil> (raw)
In-Reply-To: <200209130757.g8D7vxp09323@Port.imtp.ilyichevsk.odessa.ua>

On Friday 13 September 2002 07:53 am, Denis Vlasenko wrote:
> On 11 September 2002 16:08, Jim Sibley wrote:
> >                                     resource
> >      group                          priority                kill priority
> >      system                         0                       0 - never
> > kill support                        1                       1
> >      payroll                        2                       2
> >      production                     3                       3
> >      general user                   4                       4
> >      production backgournd          5                       3
>
>                                                              ^^^
>                                  make sure testing and general user are
> killed BEFORE production
>
> >      testing                        6                       5
>
> I like this. Maybe map it to user gid and provide /proc interface?
>
> Let's say on your server you allocated gids this way:
> 0   -   system
> 100 -   support
> 110 -   payroll
> 120 -   production
> 200 -  general user
> 130 -   production background
> 500 - testing
>
> # echo "0 100 110 120 200 130 500" >/proc/resourceprio
> # echo "0 100 110 120 130 200 500" >/proc/killprio

Don't base it on gid. Remember, a user can be a member of multiple
gids for file access. At this point you may get a payroll/production
conflict, or a production/production background conflict.

You really have to use a resource accounting structure that allows
one and only one id per process. A user may (like groups) have
access to multiple resource accounts, but a given process should
only have one.
-- 
-------------------------------------------------------------------------
Jesse I Pollard, II
Email: pollard@navo.hpc.mil

Any opinions expressed are solely my own.

  reply	other threads:[~2002-09-13 12:53 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2002-09-11 18:08 Killing/balancing processes when overcommited Jim Sibley
2002-09-11 18:27 ` Jurriaan
2002-09-12  8:26   ` Helge Hafting
2002-09-11 21:44 ` Alan Cox
2002-09-12  7:06 ` Tim Connors
2002-09-12  7:25 ` Giuliano Pochini
2002-09-12 16:02   ` Rik van Riel
2002-09-12 18:30   ` Thunder from the hill
2002-09-13  8:17     ` Giuliano Pochini
2002-09-13 10:22       ` Thunder from the hill
2002-09-13 12:53 ` Denis Vlasenko
2002-09-13 12:54   ` Jesse Pollard [this message]
  -- strict thread matches above, loose matches on Subject: below --
2002-09-12 18:14 Jim Sibley
2002-09-12 19:00 Jim Sibley
2002-09-12 19:08 ` Thunder from the hill
2002-09-12 20:35   ` Alan Cox
2002-09-12 20:43     ` Thunder from the hill
2002-09-12 20:55       ` Alan Cox
2002-09-12 21:15         ` Thunder from the hill
2002-09-12 21:22           ` Jesse Pollard
2002-09-12 23:08           ` Alan Cox
2002-09-12 23:12             ` Thunder from the hill
2002-09-12 21:19       ` Jesse Pollard
2002-09-12 21:56         ` Thunder from the hill
2002-09-13  7:51           ` Giuliano Pochini
2002-09-13 10:17             ` Thunder from the hill
2002-09-12 19:09 ` Rik van Riel
2002-09-13  8:05 ` Giuliano Pochini
2002-09-13 10:54   ` Helge Hafting
2002-09-13 13:02     ` Giuliano Pochini
2002-09-13 16:40       ` Gerhard Mack
2002-09-13 20:23   ` Timothy D. Witham
2002-09-13 21:13 Jim Sibley
2002-09-13 22:31 ` Timothy D. Witham
2002-09-13 22:38   ` Timothy D. Witham
2002-09-13 22:44   ` Rik van Riel
2002-09-13 23:12     ` Timothy D. Witham
2002-09-16  7:29     ` Helge Hafting
2002-09-16 14:03       ` Rik van Riel
2002-09-16 18:49         ` Timothy D. Witham
2002-09-16 19:11           ` Rik van Riel
2002-09-16 20:27             ` Timothy D. Witham
2002-09-14  0:23   ` Jim Sibley

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=200209130754.21751.pollard@admin.navo.hpc.mil \
    --to=pollard@admin.navo.hpc.mil \
    --cc=jlsibley@us.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=ltc@linux.ibm.com \
    --cc=riel@conectiva.com.br \
    --cc=tdreed@us.ibm.com \
    --cc=vda@port.imtp.ilyichevsk.odessa.ua \
    /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