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

Re: [Pulp-list] Pulp with minimum complexity

On Fri, Jul 5, 2013 at 2:42 PM, Michael Hrivnak <mhrivnak redhat com> wrote:
> Mongo is a hard requirement. Pulp cannot function without it. However, it requires no configuration on your part.

Ok. That's valuable info. Keep in mind that a careful sysadmin (or
infra architect) will go "omg, _another_ database that we need to know
how to backup, restore and unf**k, why can't they use
$some_db_we_know, or $same_db_the_rest_of_the_stack_uses".

In my case, it's "why aren't they using PostgreSQL, or SQLite?" (both
of which I know in depth, and I know how to back them up, and that
they don't get trivially corrupted).

There's a sad history of broken or corruption-prone backend DBs
(OpenLDAP is the poster child here, even today it ships with pretty
brittle defaults).

> qpid is theoretically optional, for now

Heh. Look, I have nothing against messaging queues, but I worked with
ejabberd (meaning, I have maintaintained feature patches, fixed up a
discreet pile of subtle bugs in it). I don't look forward to any
situation where you'd have to debug it.

> boost isn't actually a dependency.

Oh, that's great news.

> In general, I think pulp can be a good fit for a simple use case like yours.


I will be honest -- I appreciate the help and frankness. For the time
being, I am looking at very light scripting of reposync.

Will revisit Pulp as my needs evolve. If I could have a nice Pulp-y
CLI to reposync, couple cronjobs to keep it current, and backed by a
trivial config file or sqlite DB, I'd be in heaven.


 martin langhoff gmail com
 -  ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 ~ http://docs.moodle.org/en/User:Martin_Langhoff

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