[Spacewalk-list] Low bandwidth locations

[Matthew Madey]

We currently have 4 Spacewalk Proxies, geographically positioned to support
the clients. Problem is, we have 8000+ clients, and they are split up
between hundreds of locations. It's a huge environment, but we've made it
work pretty well so far. To work around the issue of bandwidth, we've been
pre-staging patch bundles by delivering the metadata and packages to the
/var/cache/yum directory on all the clients, and setting the metadata
expire in yum.conf to about 7 days. We run into trouble though when someone
runs a yum clean all on a client, and it has to download the metadata over
the slow connection. Or when we have to do zero day patching, which luckily
the most recent one's (SSL and BASH) have been very small updates. We
created a separate channel for that sort of thing, so the amount of
metadata downloaded is very small.

The traffic shaping we've done on the network layer seems to have helped a
bit, but I'm not sure what other settings we can tune in Spacewalk to
better throttle client requests. We've thought about turning off rhnsd on
the clients during peak business hours, but I'd rather not go that route.

On Tue, Nov 18, 2014 at 12:18 PM, Waldirio Manhães Pinheiro <
waldirio at gmail.com> wrote:

> Matthew,
>
> I believe be interesting know better your environment, btw, you can work
> with proxy (between your company and filial/client) or really use Spacewalk
> Proxy. With this you will probably work better, you have to check the
> necessity according the quantity of clients or the problem caused by
> latency.
>
> Do a test with one proxy in your problematic environment and enjoy!
>
> Let me know if works for you.
>
> B'Regards
>
> ______________
> Atenciosamente
> Waldirio
> msn: waldirio at gmail.com
> Skype: waldirio
> Site: www.waldirio.com.br
> Blog: blog.waldirio.com.br
> LinkedIn: http://br.linkedin.com/pub/waldirio-pinheiro/22/b21/646
> PGP: www.waldirio.com.br/public.html
>
> On Tue, Nov 18, 2014 at 2:20 PM, Matthew Madey <mattmadey at gmail.com>
> wrote:
>
>> All,
>>
>> I've run into a situation where my total metadata for channels has grown
>> quite large (about 200MB). I've done all I can as far as channel\package
>> management to reduce the size of the metadata..
>> We have a number of locations where the bandwidth to the clients is
>> extremely small.. So downloading new metadata causes severe network latency
>> on these clients.. We've done some traffic shaping on the network layer to
>> reduce the impact, but I'm looking for other alternatives focused more
>> around Spacewalk. Looking for anyone who may have a similar situation and
>> what steps they've taken to mitigate the issue. Are there any sort of
>> throttling settings available in Spacewalk, or does this all have to be
>> done at the OS\Network layers?
>>
>>
>>
>> _______________________________________________
>> Spacewalk-list mailing list
>> Spacewalk-list at redhat.com
>> https://www.redhat.com/mailman/listinfo/spacewalk-list
>>
>
>
> _______________________________________________
> Spacewalk-list mailing list
> Spacewalk-list at redhat.com
> https://www.redhat.com/mailman/listinfo/spacewalk-list
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/spacewalk-list/attachments/20141119/9a9fb528/attachment.htm>