yum update so slow

[Jochen Schlick]

James Antill wrote:
> On Mon, 2007-09-24 at 11:45 +0200, Jochen Schlick wrote:
>> Hi, I have a laptop where rawhide is installed since the days
>> of Fedora Core 1. It has only 256MB ram so updating using yum
>> was always like a lame duck but nowadays it's really a crap.
> 
>  This might not help, but if you'd like to try something I've attached
> some python code which calls yum and "iterativlely updates". This might
> make it be small enough that it isn't living in swap, or at least not as
> badly.
>  Let me know how it goes.
> 

I tried yesterday evening it on my laptop and on a ram-shrinked(256MB) rawhide
qemu/kvm-guest at home. At first it looked fine, but when the script started
to update with X11-stuff (perhaps due to their dependencies or so) it failed to
update with "Error: Failed to update:..."  the script tried the next one
and so on. After an hour I realized that both systems were no longer reachable
and continuously swapping.
This morning I have the following results:
- The laptop was dead and after a reboot it hangs (so it needs further forensic
   analysis ;-)
- The rawhide quemu/kvm guest is alive, but most of the 70 updates are failed
   when I believe the update script. /var/log/messages shows a lot of oom-killer
   activity. On first sight I think every failed update corresponds with at least
   one oom-killer entry...


best regards


-- 
---------------------------------------------------------------------------
Jochen Schlick
---------------------------------------------------------------------------