[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: [Pulp-list] Clustering and HA



Matthew,

I saw you mentioned clustering/high availability in the title, but scaling out Pulp to handle more traffic in the email. It sounds like you are on the right track, but I just wanted to confirm you are trying to solve for handling increased load and not to maintain service availability during hardware/OS failure. Both are possible, but it's easier to solve one problem first and then tackle the other.

On Mon, Jul 23, 2018 at 8:16 AM, Brian Bouterse <bbouters redhat com> wrote:
Hi Matthew,

There are many users who scale Pulp to those levels and several installations that have 50K+ clients. You can horizontally scale Apache and/or the Pulp workers to get the throughput and/or availability you need. You could do it all in containers, all on VMs, bare metal, or any mixture. In terms of docs, have you seen the scaling guide? If not, it's here:  https://docs.pulpproject.org/user-guide/scaling.html If you run into issues, feel free to email the list with other questions.

Also, we invite community blog posts if you want to share with others on how you're deploying/using Pulp. Here are some docs on how to contribute that, if that's interesting to you: https://pulpproject.org/2016/12/05/contributing-blog-post-to-pulpproject.org/

Cheers,
Brian


On Sun, Jul 22, 2018 at 12:55 PM, Matthew Madey <mattmadey gmail com> wrote:
I'm looking to do a proof of concept Pulp cluster to see if it can meet the needs of a large enterprise. I've been looking through the documentation on the Pulp website, but was wondering if anyone has done this at scale and has come across any more in depth documentation than what's on the official website? My scenario would be a very large amount of clients (20,000+), with very active repositories from developers pushing packages (roughly 3000-5000 new RPM's a week). I'm thinking apache and the pulp workers will be two of my biggest concerns when scaling. I'm even thinking of possibly a hybrid scenario with the workers being containers, possibly even the apache servers as well. This would allow me to scale horizontally very quickly if needed. Any input would be helpful! Thanks!

_______________________________________________
Pulp-list mailing list
Pulp-list redhat com
https://www.redhat.com/mailman/listinfo/pulp-list


_______________________________________________
Pulp-list mailing list
Pulp-list redhat com
https://www.redhat.com/mailman/listinfo/pulp-list


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]