Hi!
On 08/22/2009 01:56 PM, Linux Advocate wrote:
> guys, i have the yum plugin -> fastest mirror . But not even once i have seen it selecting repos which are near my region such as japan or australia ( where i get the best speeds). Something is wrong.
There are a few different things that are at play here. Firstly, the
mirrorlist will do a geoip lookup and give you what should be the 'best
list' for you, so its possible the ip you are connecting from isnt being
recognised as being in Japan. from the list you posted, it clearly looks
like its passing you European mirrors. So that would be the first place
to test.
Could you open an issue report at bugs.centos.org against the 'Mirror'
project, with what ip / subnet you are connecting from ?
> It seems stuck with these 3 sites ;
>
> Determining fastest mirrors
> * ftp-stud.fht-esslingen.de : 0.309373 secs
> * apt.sw.be : 0.483867 secs
> * fr2.rpmfind.net : 0.503842 secs
So, there are a few options you have here - you can either just replace
the mirrorlist with the baseurl= line in /etc/yum.repos.d/ - however, I
would strongly encourage you to NOT do that. You are much better off
resolving the issue at hand. If you look at
/etc/yum/pluginconf.d/fastestmirror.conf you will see the default
settings, which should be something like this :
[main]
enabled=1
verbose=0
socket_timeout=3
hostfilepath=/var/cache/yum/timedhosts.txt
maxhostfileage=10
maxthreads=15
in order to refresh the cash, 'yum clean all' should remove
/var/cache/yum/timedhosts.txt ; however if thats failing, another issue
report on bugs.centos.org! also try to remove it manually - its possible
some acl layer is getting in the way.
Finally, the mashostfileage ( in days )[1] - you can consider reducing
that to '1'. During the first few days ( < 5 ) from the time a new
release goes out ( like 4.8 ), mirror performance can vary by large
factors on a daily basis. It might be a good idea to reduce that on your
machine anyway, and see if that builds your confidence in the way this
works.
Hope this helps.
[2] its not hard reducing the resolution of mashostfileage, I plan on
investigating what the implications around it might be.
--
Karanbir Singh : http://www.karan.org/ : 2522219@icq
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
http://lists.centos.org/mailman/listinfo/centos
