Hi all,
I am getting some pressure to move fedocal to production (which is pretty nice).
So just to be sure the next steps should be: - roll out the 0.1.0 release - package it - it now has a setup.py so I hope that it'll be pretty simple - request for hosting/hw ? - create the adequate puppet module - which url? admin.fedoraproject.org/calendar ?
Do I miss anything?
Also, what's the traditional time-line to move from stg to prod?
Thanks, Pierre
On Tue, 12 Feb 2013 19:42:56 +0100 Pierre-Yves Chibon pingou@pingoured.fr wrote:
Hi all,
I am getting some pressure to move fedocal to production (which is pretty nice).
So just to be sure the next steps should be:
- roll out the 0.1.0 release
- package it
- it now has a setup.py so I hope that it'll be pretty simple
- request for hosting/hw ?
- create the adequate puppet module
- which url? admin.fedoraproject.org/calendar ?
Sounds reasonable. In addition:
- What are the monitoring requirements?
- Should we use caching/proxies for it?
- If it gets a lot of load, can we add more instances and load balance them?
- Whats the db requirement?
Do I miss anything?
I think admin.fedoraproject.org/calendar or perhaps apps.fedoraproject.org/calendar would be fine.
Also, what's the traditional time-line to move from stg to prod?
Mostly stg is so we can get puppet sorted out and find any issues with proxying and caching and such. It can sometimes be pretty short if there's no particular issues found.
kevin
On Tue, 2013-02-12 at 13:00 -0700, Kevin Fenzi wrote:
On Tue, 12 Feb 2013 19:42:56 +0100 Pierre-Yves Chibon pingou@pingoured.fr wrote:
Hi all,
I am getting some pressure to move fedocal to production (which is pretty nice).
So just to be sure the next steps should be:
- roll out the 0.1.0 release
- package it
- it now has a setup.py so I hope that it'll be pretty simple
- request for hosting/hw ?
- create the adequate puppet module
- which url? admin.fedoraproject.org/calendar ?
Sounds reasonable. In addition:
- What are the monitoring requirements?
None specific, the standard package?
- Should we use caching/proxies for it?
Caching might be nice, especially for the ical feed if lot of people subscribe to it/them.
- If it gets a lot of load, can we add more instances and load balance them?
Toshio said that should work for flask. I did not perform any tests though.
- Whats the db requirement?
Postgres is fine.
Do I miss anything?
I think admin.fedoraproject.org/calendar or perhaps apps.fedoraproject.org/calendar would be fine.
hm, since it needs login shouldn't it be admin?
Also, what's the traditional time-line to move from stg to prod?
Mostly stg is so we can get puppet sorted out and find any issues with proxying and caching and such. It can sometimes be pretty short if there's no particular issues found.
Cool! :)
Thanks for the feedback,
Pierre
On Tue, Feb 12, 2013 at 09:15:42PM +0100, Pierre-Yves Chibon wrote:
On Tue, 2013-02-12 at 13:00 -0700, Kevin Fenzi wrote:
On Tue, 12 Feb 2013 19:42:56 +0100 Pierre-Yves Chibon pingou@pingoured.fr wrote:
Hi all,
I am getting some pressure to move fedocal to production (which is pretty nice).
Good! :)
I think admin.fedoraproject.org/calendar or perhaps apps.fedoraproject.org/calendar would be fine.
hm, since it needs login shouldn't it be admin?
I'm not sure what the original reason for the distinction was, but fyi apps.fedoraproject.org/tagger requires login.
If it's just an aesthetic choice, I prefer the apps.fp.o/calendar one. Either way, I can add this to the apps.fp.o/ landing page when its ready.
-Ralph
On Tue, 12 Feb 2013 18:34:31 -0500 Ralph Bean rbean@redhat.com wrote:
I think admin.fedoraproject.org/calendar or perhaps apps.fedoraproject.org/calendar would be fine.
hm, since it needs login shouldn't it be admin?
I'm not sure what the original reason for the distinction was, but fyi apps.fedoraproject.org/tagger requires login.
Originally, admin.fedoraproject.org was setup as https only and applications under there could share cookies, so you could login to say bodhi and get automatically logged into pkgdb.
When we reorganized things a while back the thought was that 'admin.fedoraproject.org' confuses people because it looks like some kind of administrative site, not an end user facing application.
So, we have been moving to move applications to 'apps.fedoraproject.org'.
If it's just an aesthetic choice, I prefer the apps.fp.o/calendar one. Either way, I can add this to the apps.fp.o/ landing page when its ready.
Cool.
kevin
On Tue, Feb 12, 2013 at 01:00:59PM -0700, Kevin Fenzi wrote:
On Tue, 12 Feb 2013 19:42:56 +0100 Pierre-Yves Chibon pingou@pingoured.fr wrote:
Hi all,
I am getting some pressure to move fedocal to production (which is pretty nice).
So just to be sure the next steps should be:
- roll out the 0.1.0 release
- package it
- it now has a setup.py so I hope that it'll be pretty simple
Ok the part above is done, review is up at: https://bugzilla.redhat.com/show_bug.cgi?id
- request for hosting/hw ?
I will get started on this
- create the adequate puppet module
Puppet or ansible? Which basically raise the question, what more do we need to start moving apps to Ansible? I kinda think it would make sense for new apps to be deployed via ansible rather than puppet since we want to switch anyway, but that of course rely on what is needed still :)
Pierre
On Sun, 24 Feb 2013 17:13:40 +0100 Pierre-Yves Chibon pingou@pingoured.fr wrote:
On Tue, Feb 12, 2013 at 01:00:59PM -0700, Kevin Fenzi wrote:
On Tue, 12 Feb 2013 19:42:56 +0100 Pierre-Yves Chibon pingou@pingoured.fr wrote:
Hi all,
I am getting some pressure to move fedocal to production (which is pretty nice).
So just to be sure the next steps should be:
- roll out the 0.1.0 release
- package it
- it now has a setup.py so I hope that it'll be pretty simple
Ok the part above is done, review is up at: https://bugzilla.redhat.com/show_bug.cgi?id
You mean: https://bugzilla.redhat.com/show_bug.cgi?id=915074 :)
...snip...
Puppet or ansible? Which basically raise the question, what more do we need to start moving apps to Ansible? I kinda think it would make sense for new apps to be deployed via ansible rather than puppet since we want to switch anyway, but that of course rely on what is needed still :)
Sure, we still need a few things:
1) Need to figure out at least a good starting point for organizing things. We can always adjust, but I'd like to think about our common uses like staging and come up with something to start with that can work.
2) We need periodic cron jobs setup. Possibly we have it fire only when there are changes in the repo + some wait time.
3) convert the sane parts of global and other 'base' stuff we do on all machines to ansible from puppet. I think we can drop a bunch of useless stuff, and also adjust our kickstarts to just install all the stuff we want so we don't need to adjust packages and such.
4) To be paranoid (this applies mostly to migrating hosts from puppet) I would like to: a) convert a host to ansible. b) run both puppet/ansible in turn on it and make sure things don't flip/flop. c) re-install fresh and run ansible, then run puppet and make sure it doesn't do anything that we want/need.
That way we can make sure changes are sane.
I think 1-3 are what we need before new hosts in ansbile. 4 is needed after that after we start converting things over.
kevin
infrastructure@lists.fedoraproject.org