|operations/puppet : production||Allow override of sync_common config|
|Declined||None||T88701 [Quarterly Success Metric] Green nightly builds on the staging cluster (tracking)|
|Declined||None||T88702 Create staging cluster|
|Declined||None||T91548 Create staging-mw-app* (MW App servers)|
Most of the role::mediawiki::appserver seems to work without alteration.
One problem is sync_common running in ::mediawiki::scap.
Scap's ini config assumes anything running in wmflabs should use deployment-bastion.eqiad.wmflabs as the master_rsync. There seem to be a couple of problems with this, the largest of which is either (a) This won't work for testing deploys in staging or (b):
thcipriani@staging-mw01:~$ nc -vz deployment-bastion.eqiad.wmflabs -w 1 873 nc: connect to deployment-bastion.eqiad.wmflabs port 873 (tcp) timed out: Operation now in progress
There was a local hack on staging-tin to deal with this: setting an rsync_master for the hostname staging-tin.eqiad.wmflabs in the scap.cfg ini
We'll need to pass this info in via hiera, I'd expect.
Got some warnings setting up first app server. I think most of this will go away with T1256 :
- wikispecies.conf – live handles this in redirect.conf and remnant.conf which ultimately handles it via wikimedia-common.incl → DocumentRoot "/srv/mediawiki/docroot/wikimedia.org"—main problem in staging is that there is no "/srv/mediawiki/docroot/wikispecies.org" which is the current specified docroot in beta
- /mnt/upload7 – doesn't currently exist on any beta instance(?) is defined as the docroot of upload.beta.wmflabs.org, this will be replaced by swift I'm assuming
- /srv/mediawiki/docroot/config – Also doesn't currently exist on any beta instance. Docroot of config.beta.wmflabs.org
- /data/project/logs – defined in beta/logging.conf