[OFFLIST] Re: Ubuntu

Mark Constable markc at renta.net
Thu Jan 8 03:39:12 UTC 2015


FWIW, and offlist because I am being critical of Arch, and Debian.

> The reason we were asking about Arch instructions, is because it's
> already been packaged there by some nice folk.

I was briefly a maintainer of the original AUR package. I didn't do much
but kept it on life support until someone else more competent took over.

> We're looking for someone to get a PPA started for Ubuntu, or at least
> get one started, and then I could perhaps contribute to it. I agree that
> it should be pretty straightforward.

Yes, for anyone that has already created PPA packages it's probably a
breeze but for me I'd have to start from scratch and RTFM and compared
to using makepkg on Arch I find it hard to get enthused.

This is the best I can do atm (spread the word a bit)...

https://github.com/oerdnj/deb.sury.org/issues/38

Perhaps if you add a comment it might help Ondre consider it.

> Are you interested in working together on an Ubuntu PPA?

I certainly could provide some testing.


***

This is the real FWIW part, just ignore if busy...

I gave up on Arch because of their abrupt switch to systemd which broke
my Arch servers for the 3rd time in a year (after a move to /usr/lib and
all binaries in /usr/bin). I can deal with that but I could never let
my clients use Arch on their servers... so being pee'd off with Debian
for being too slow with updates I settled on ubuntu as my best choice.

Now I have nginx 1.6, php 5.5, mariadb 5.5 on all servers and even the
latest courier-mta thanks to Ondres PPA. Me now 99.9% happy.

The only reason I am at all interested in systemd is because of cockpit.

I'm not against systemd like the Devuan folks but I am wary that such a
significant change to major core systems is... yet to be proven (to me)
as being a worthwhile change and worthy of the code bloat.

However, a websocket server in front of a system, like systemd, suddenly
makes sense of systemd and definitely justifies the monolithic approach
of systemd, and even the use of Go to build cockpit itself is also
justified for performance (compared to a scripted language.)

My main interest with cockpit, once I can use it on a current ubuntu
server, will be to provide an ultra simple PHP+BS4 frontend so the
cockpit workflow can be more easily integrated with other PHP apps
on a standard LEMP server.



More information about the cockpit-devel mailing list