[sr-dev] RFC: infrastructure upgrade - git, tracker, ...

[Daniel-Constantin Mierla]

Hello,

as most of you know, we have a rather distributed infrastructure, with
servers provided from different companies or persons.

We came to the time when one of the servers is too old and considered to
be decommisioned, so we have to decide how to move on. It is about
sip-router.org, who was offered/sponsored by courtesy of Jan Janak. From
project point of view, the server hosts:

- git repository
- bug tracker
- website and wiki under domain sip-router.org

Given that no matter what we like, there is work to do, I am looking to
see what are the best options out there for everyone involved in the
project.

The sip-router.or wbesite and wiki, which are not really updated anyhow,
will be relocated as virtual host in kamailio.org server and made static
for historic purposes.

For git and tracker, I thought of two variants:

1) move to use an external hosting service - the first candidate and
perhaps the only to be considered is GitHub, we have there already a
real time mirror of git repository. Then we should get a read only
mirror to kamailio.org. If the tracker on github is good enough for
everyone, then we will use it. I could see quickly that lot of kamailio
developers already have an account on github.

2) get a new server and relocate those components there. It will need to
be configured from scratch and the components eventually updated to use
latest versions. In case of tracker, we have eventually to re-evaluate
if flyspray worth keeping, as we had several discussions, due to the
fact that the project doesn't seem to be very active.

My personal preference at this moment is 1), given that offloads
administration works from project.

For 2) we will need someone to commit to a sysadmin job for long time.
As probably you noticed lately, serious security vulnerabilities can
appear and someone needs to take care of proper maintenance of the
server. I don't want to get all services on kamailio.org, as it has
other critical components (mailing lists, releases, website, ...) and
messing it or overloading doesn't make sense.

There is no real pressure to come to a decision, we can still rely on
the server for a while, but I would rather not postpone it for long.

While users are encouraged to give their opinion, I feel that existing
developers should have the main role in decision, being something that
impacts them directly.

Your preference? Any other opinions?

Cheers,
Daniel

-- 
Daniel-Constantin Mierla
http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda
Kamailio Advanced Training, Nov 24-27, Berlin - http://www.asipto.com