[Spacewalk-list] proxy issue - filelists.xml.gz corrupt?

[James Hogarth]

On 14 July 2010 12:21, Tom Brown <tom at ng23.net> wrote:
> On 14 July 2010 00:46, James Hogarth <james.hogarth at gmail.com> wrote:
>> It's bedtime for me here now but search the mailing lists for my posts about
>> squid configuration.  Tomorrow noon ish bst I'll provide more detailed info
>> for you if you need it but you are on the right track.
>
> thanks for this - i see this in the archives
>
> http://markmail.org/thread/djtb66pjsumfc4sn#query:+page:1+mid:ehbctr2h4afpribb+state:results
>
> i wonder if that resolved the issue you were seeing as i feel its a
> similar issue for me
>
> thanks
>
> _______________________________________________
> Spacewalk-list mailing list
> Spacewalk-list at redhat.com
> https://www.redhat.com/mailman/listinfo/spacewalk-list
>

Our 'production' repo sits somewhat outside of the envelope expected on this...

We release code to production systems once per week for our
application and with the tweak done we were still seeing out of sync
behaviour of repodata, This is what an excerpt from my squid.conf
looks like on my proxies for our production environments now:

refresh_pattern  \.rpm$  10080 100% 525960 override-expire
override-lastmod ignore-reload reload-into-ims
refresh_pattern    /XMLRPC/GET-REQ/skyentprod/repodata/.*\.xml.* 0 1% 15
refresh_pattern    /XMLRPC/GET-REQ/.*/repodata/.*\.xml.*$ 0 1% 525960
refresh_pattern 	.		0	100%	525960

As you can see we took the change that was added before but then added
a line before it to explicitly match the URI for our custom production
repo with a max lifetime of 15 minutes. This repo only changes once a
week and we have found this configuration optimised for our setup...
the 1% is not modified checks didn't work nicely for us on this due to
the long changes on updates...

Adjust as required and your mileage may vary ^^

James