break freeze request: increase MM crawler timeout
Toshio Kuratomi
a.badger at gmail.com
Sat Dec 15 15:29:08 UTC 2012
+1
Better to try this now and iron out any problems than later. You'll be
around to tweak further/revert if necessary?
-Toshio
On Dec 15, 2012 6:24 AM, "Matt Domsch" <Matt_Domsch at dell.com> wrote:
> The mirrormanager crawler has a 60 minute timeout - if it can't crawl
> the mirror in 60 minutes, it assumes it's running slow enough not to
> be able to serve traffic reasonably, and marks it as not up-to-date,
> dropping it from the mirror lists.
>
> With the new alphabetical Packages/[0-9a-z]/*.rpm directory structure,
> the crawler now as an additional ~36 directories to search in each of
> 24 (currently) directory trees, at 10 RPMs each being searched, at ~6
> RPMs/second being checked, is another 24 minutes more than was
> previously needed. This is pushing mirrors that were previously
> completing in just over 30 minutes, over the timeout.
>
> I'd like to increase the timeout from 60 to 90 minutes. I would like
> to change the number of RPMs being looked at in each directory from 10
> to 3 (it's the most recent 10 vs most recent 3; if you have the most
> recent 3, you will have the next most recent 7 I would think...)
>
> Thanks,
> Matt
>
> --
> Matt Domsch
> Technology Strategist
> Dell | Office of the CTO
> _______________________________________________
> infrastructure mailing list
> infrastructure at lists.fedoraproject.org
> https://admin.fedoraproject.org/mailman/listinfo/infrastructure
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.fedoraproject.org/pipermail/infrastructure/attachments/20121215/36c915f4/attachment.html>
More information about the infrastructure
mailing list