[Mailman-Users] Re: qrunner with over 90% CPU

Paul H Byerly paul at thcwd.com
Tue Aug 12 03:12:45 CEST 2003

Jon Carnes  wrote:
>Sorry to pepper you with so many questions..

      Hey, anything that might help.

>What do you get when you type:
>   cat /proc/sys/fs/file-max


>   cat /proc/sys/kernel/threads-max


>On one of my well-running RH 9 boxes,

      I'm on 7.1.

>  I get:
>   cat /proc/sys/fs/file-max
>     26208

      I'm more than double that.  Is more too many?

>   cat /proc/sys/kernel/threads-max
>     4095

      I'm very close to that.

>You can adjust these values by becoming root and typing in something like:
>   echo 16384 > /proc/sys/fs/file-max
>   echo 2048 > /proc/sys/kernel/threads-max
>Note: this is not a recommendation (well it is, but it's not *my* 

      Both of which are well less than what I have.  What would this do?

>BTW: I'm guessing that you are not running either NFS or NIS (NIS is a way 
>of sharing user information amoung a lot of unix servers)

      Sounds right.

>I also guess that hda1 is your /boot and that hda3 is /


>That indicates that your install is using one IDE drive.

      Right again.

>I hope your mail traffic is not too high!

      Not yet, but it will be in time as I move more lists.  I am way over 
due for some upgrades and the box no longer covers what I need, so I'm 
looking to move to a new server.  I'm thinking a two HD box running 
C-panel, and I'm looking to run a better MTA than sendmail.

>You might want to drop the Open Relay checks for a week and see if that
>makes a difference. I'm not sure, but Sendmail may be holding open the
>access.db file while it does these checks.

      Could be, but my lists are currently announcement lists that send at 
a low mail time.  I once let the runaway process go for 15 minutes to see 
if it would clear, it did not.  The open relay checks cut my mail volume 
substantially, and I suspect I'd hear from some customers if I stop them.

      Thanks for the help.

<>< Paul 

More information about the Mailman-Users mailing list