On 16 February 2018 at 14:20, Stephen John Smoogen <smooge@xxxxxxxxx> wrote:
> On 16 February 2018 at 08:36, Günther J. Niederwimmer <gjn@xxxxxxxxxxx> wrote:
>> Hello,
>> I have thousands of this messages on my servers ??
>>
>> Is this a Problem on my site or is the infrastructure broken??
>> /etc/cron.hourly/0yum-hourly.cron:
>>
>> Could not get metalink https://mirrors.fedoraproject.org/metalink?
>> repo=epel-7&arch=x86_64 error was
>> 14: HTTPS Error 503 - Service Unavailable
>>
>> Thanks for a answer,
>>
>
> Hello, from the Fedora side of Infrastructure. We have been trying to
> get the 503 error rate down in the last six months and thought we had
> licked it with recent changes. We had moved down the error rate to
> 80,000 bad requests per day out of 13.5 million total ones which
> actually is a creep up from what we had last month. Most of the
> requests every day happen in the time frame of XX:00 -> XX:15 of every
> hour. Every proxy seems to give about 100 per day but 2 proxies seem
> to have had docker problems and have been each contributing 32,000 to
> 38,000 503's.
>
> We did find an error in the docker restart script which hopefully will
> drop this further. We are still evaluating why just those 2 proxies
> have so many more and will let you know what is found.
>
> Thank you for your patience.
>
To follow up we have cut the number of 503's down to 1200 per 13.5
million per day. If you are still seeing problems with EPEL updates
please let us know with a time stamp and ip address of when they are
occurring. I am trying to track down if we have a problem with
specific proxies or times.
--
Stephen J Smoogen.
_______________________________________________
CentOS mailing list
CentOS@xxxxxxxxxx
https://lists.centos.org/mailman/listinfo/centos
