On Tue, Apr 20, 2004, seth vidal <skvidal@xxxxxxxxxxxx> wrote:
> On Tue, 2004-04-20 at 15:29 -0700, Johannes Erdfelt wrote:
>
> > I'm not sure if I haven't noticed it until recently, but yum is a pig
> > when it comes to memory utilization:
> >
> > root 17057 99.9 45.1 1534452 936384 pts/3 R 15:20 9:01
> > /usr/bin/python /usr/bin/yum update
> >
> > As you can see, the process is 1.5GB in size with 936MB resident. This
> > is on a system with 2GB of memory.
> >
> > On a system with 256MB of memory that I have, it spends much more time
> > swapping than actually doing an update of my system.
> >
> > It has come to the point where I feel I need to do something about this
> > since it's just not acceptable to me.
> >
> > Has anyone dug into this any deeper? Is it the RPM libraries? Or just
> > the algorithm that yum uses?
>
> read the archives.
Any pointers? I went back quite a few months and didn't see anything
related.
> it's the depresolution using addInstall('a') modes.
>
> It's not going to be used in the future but changing the depresolution
> mechanism in yum 2.0.x now doesn't seem to be the best use of time.
I guess it depends on how significant a problem it is then. For me, it's
enough that I want to fix it.
> in your particular case I'm betting you have a bunch of kernels
> installed. It's a well known situation that yum and rpm get very very
> large when doing updates/installs with lots of kernels installed.
Yup. I've been doing some more tests and can see that it's definately
related to the number of kernels installed.
However, it's still an insane amount of memory usage with a small number
of kernels installed.
JE
