Re: PostgreSQL uses more memory on 6.1?



On Tuesday 23 May 2006 16:05, Claus Guttesen wrote:

He probably meant the following kernel-settings:

options SHMMAXPGS=393216
options SEMMNI=240
options SEMMNS=1440
options SEMUME=240
options SEMMNU=720

Those are for my quad-opteron server with 8 GB RAM. Adjust
accordingly. Did you compile a custom-build-kernel or GENERIC?

I hadn't compiled any of those into my custom kernel, as per some now
long-forgotten web page that said I should use the sysctls to reach the
same effect.

What I'd *really* like to find is an explanation of how to set those
numbers. I'm using a single Xeon with 3GB of RAM and have no idea how high
to set those, or whether a particular value is too high. Then again, once
I figure it out I guess I should have a fair amount of job security. :-)
--
Kirk Strauser
_______________________________________________
freebsd-stable@xxxxxxxxxxx mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscribe@xxxxxxxxxxx"



Relevant Pages

  • Re: Dear ms VS.NET 2005 development team
    ... starting, maybe), there must be something wrong somewhere because 2 GB RAM ... cannot be productive waiting 3 minutes to compile a simple winform. ... rewrite big chunks of the OS in .net, ... Hopefully by the time I post this, my test winform project that i F5'd ...
    (microsoft.public.vsnet.general)
  • Re: 2.6.0-test9 - poor swap performance on low end machines
    ... > the detriment of other tasks that are on the runqueue and still need ram. ... The test compile was not of the kernel but of a file in a C++ program ... Being able to choose a manual setting for swappiness is not "futile". ...
    (Linux-Kernel)
  • Swapping in 2.6.10 and 2.6.11.11 on a desktop system
    ... I upgraded to 2.6.11.11 from 2.6.8.1 yesterday and tried to compile ... About an hour ago I started a very RAM consuming application and at 2/3 RAM ... usage I get swap hell. ... yet ram usage _and_ swap ...
    (Linux-Kernel)
  • Re: [PATCH] sched.c: Be a bit more conservative in SMP
    ... > Maybe I was completely wrong with my assumption that one running process ... > the load on one cpu and put too many processes on it, ... copy the sources to a tmpfs directory and compile ... ram would solve so that it's all compute bound. ...
    (Linux-Kernel)
  • Re: Getting forms to load faster..
    ... My development machine is a P4 2.8 GHz machine with a gig of RAM. ... compile my project and run it I can actually see the forms being drawn ...
    (microsoft.public.dotnet.languages.vb)