Hi,
Yesterday evening I added a TIMESTAMP to the root of my tree in order to follow up on the correctness of mirrors. And after 12 hours more than 50% of the mirrors do not have that TIMESTAMP file.
It is possible that it is filtered out, so the test is not at all valid. I am going to reduce the number of mirrors and only add the ones that have a recent TIMESTAMP.
However since information on why this happens (or when this happens) is scarce, I am not sure if reducing the mirrorlist will be effective.
So hopefully, after this weekend the problems some of you have seen with Yum and the mirrorlist will be a problem of the past.
Dag Wieers wrote on Fri, 12 Oct 2007 13:35:42 +0200 (CEST):
Yesterday evening I added a TIMESTAMP to the root of my tree in order to follow up on the correctness of mirrors. And after 12 hours more than 50% of the mirrors do not have that TIMESTAMP file.
It is possible that it is filtered out, so the test is not at all valid. I am going to reduce the number of mirrors and only add the ones that have a recent TIMESTAMP.
However since information on why this happens (or when this happens) is scarce, I am not sure if reducing the mirrorlist will be effective.
So hopefully, after this weekend the problems some of you have seen with Yum and the mirrorlist will be a problem of the past.
This would be great. If it still happens is there anything one could compare (except for the TIMESTAMP file)? I went to ftp://ftp.tu-chemnitz.de/pub/linux/dag/ (official apt.sw.be mirror, but there with a "wftp" hostname) and looked around, there's no TIMESTAMP file. But there is a _LASTMIRROR.txt from tonight.
Kai