[Pulp-dev] pulp3 architecture / HA

[Brian Bouterse]

Pulp3's 3 primary software components are the rest-api, content-app, and
the tasking system (its workers). Each is scalable by deploying more
processes, which can provide both increased availability and capacity. The
docs for the architecture and deployment are here:
https://docs.pulpproject.org/en/3.0/nightly/components.html

Another HA way people deploy Pulp across the WAN is to have two independant
Pulp's and have them sync to/from each other over the WAN periodically. In
this way people avoid running their database connections and binary data
over the WAN. Also Pulp3 integrates with S3 to reduce the amount of WAN
traffic, making it really just the DB traffic, which I wouldn't recommend
running over a WAN link.

At this time Pulp3 should be fully HA for failover except for 1
sub-component of the tasking system (the resource manager). That gap is
tracked here ( https://pulp.plan.io/issues/3707 ). We have a plan to solve
it, but we haven't implemented it yet.

We invite user questions through pulp-list at redhat.com, if you'd be willing
to ask through that mailing list next time, we can respond through there
also. More questions are welcome!

Cheers,
Brian


On Tue, Oct 1, 2019 at 4:15 AM <Felix.Fuhrmann at itzbund.de> wrote:

> Hello Pulp-Team,
> I have some architecture question.
> If I use pulp3 is there a possibility  to have two or three "pulp3-nodes"
> for a higher availability?
>
> Regards,
> Felix
>
>
>
>
>
>
> _______________________________________________
> Pulp-dev mailing list
> Pulp-dev at redhat.com
> https://www.redhat.com/mailman/listinfo/pulp-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/pulp-dev/attachments/20191001/a5ce0567/attachment.htm>