Freeze Break request: playbook run over all hosts with fedmsgdconfig tag
Pierre-Yves Chibon
pingou at pingoured.fr
Wed Jul 29 15:29:08 UTC 2015
On Wed, Jul 29, 2015 at 09:22:56AM -0600, Kevin Fenzi wrote:
> Greetings.
>
> The addition of mm-frontend02 yesterday meant that all the dynamically
> generated fedmsg config needs to add a entry for it.
>
> This means the check/diff daily report is pretty large because all
> these are pending.
>
> I would like to run the master playbook with -t fedmsgdconfig and then
> -t fedmsgmonitor
>
> I don't think there will be any issues.
>
> +1s?
>
> This will make the following changes on fedmsg enabled hosts:
>
> --- before: /etc/fedmsg.d/endpoints.py
> +++ after: /srv/web/infra/ansible/roles/fedmsg/base/templates/endpoints.py.j2
> @@ -41,20 +41,24 @@
> ],
>
> "askbot.ask02": [
> "tcp://ask02.%s:30%02i" % (suffix, i)
> for i in range(6)
> ],
>
> # The mirrormanager2 frontend runs as 2 processes with 1 thread each.
> "mirrormanager2.mm-frontend01": [
> "tcp://mm-frontend01.%s:30%02i" % (suffix, i)
> + for i in range(2)
> + ],
> + "mirrormanager2.mm-frontend02": [
> + "tcp://mm-frontend02.%s:30%02i" % (suffix, i)
> for i in range(2)
> ],
>
> # mizdebsk says he thinks we'll need four sockets. Two services
> # (koschei-resolver and koschei-polling) are only sending messages, one
> # service (koschei-watcher) can both send and receive them. The wsgi
> # webapp doesn't use fedmsg at all and all services are single-threaded
> # koschei-watcher uses two endpoints
> "koschei.koschei01": [
> "tcp://koschei01.%s:30%02i" % (suffix, i)
>
+1 for me
Pierre
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 181 bytes
Desc: not available
URL: <http://lists.fedoraproject.org/pipermail/infrastructure/attachments/20150729/fc90b784/attachment.sig>
More information about the infrastructure
mailing list