On Thu, Oct 17, 2013 at 06:15:57PM -0400, R P Herrold wrote: You are absolutely right that we're lacking in the documentation. Part of it is because we want to change how it's done now, but a short walk through before I go away for a couple of days. First of all, you clone the repo (and you got the correct URL): git clone --recursive git://gerrit.ovirt.org/infra-puppet.git Then you make the changes as you normally would. I recommend creating topic branches. When you've committed (with sign-off), you should submit the patches to gerrit. I highly recommend git-review for this, but you already figured this out for vdsm. Then it should be reviewed. Once it's merged, comes the tricky part. The deployment to production is manual. Trivial, but manual. What I've done, is adding a remote production git remote add production puppet@foreman.ovirt.org:puppet git fetch production # To get the references Then check out the production branch, merge master and push it git checkout production # Should base on production/production git merge master git push I should note that currently only my SSH key is (manually) added to puppet(a)foreman.ovirt.org/~puppet/.ssh/authorized_keys. Since dcaro has +2 permissions, I'd recommend him to add his until we sort out the r10k deployment (See http://gerrit.ovirt.org/19141 for that). For actually writing the patches, I'm afraid we have little to nothing in place to make your life easier. There is no testing environment, no rspec-puppet based unit tests, no rspec-system based system tests. Not even a simple rake task to validate the syntax[1]. What I can recommend you is puppet-lint and puppet parser validate. If you're a vim user, syntastic (https://github.com/scrooloose/syntastic) does this for you if it's available every time you save. If this is lacking any info or something is unclear, please do ask. I should be back monday, but I can't promise how late or if I'll make the meeting. [1]: If you are going to write tests, I did publish a WIP module which has very little classes and should be a good example: https://github.com/ekohl/puppet-profile_foreman

On Fri, 18 Oct 2013, Dan Kenigsberg wrote: After the mailing list mention of slow connectivity, and the outage on www.ovirt.org, I set up nagios monitoring to 'watch' all the principal oVirt resources. I am pretty regularly getting 'slow ping' reports as to jenkins.ovirt.org ... but on none of the others [herrold@centos-6 ~]$ !1360 ssh -l root monitor.lsbtest.org Last login: Thu Oct 17 11:52:30 2013 from (elided) [root@monitor ~]# cd /etc/nagios/objects [root@monitor objects]# grep jenkins ovirt.cfg members www.ovirt.org, gerrit.ovirt.org, \ jenkins.ovirt.org, resources.ovirt.org, lists.ovirt.org [root@monitor objects]# Should I be 'watching others'? Perhaps a ticket should be opened with the provider for that unit? -- Russ herrold

last i understood from investigation done by kiril & apuimedo, it was related to outdates python-nose pkg. which was updated via pip to latest. can you return your test? also, danken as we talked, might worth increasing the timeout or removing unit tests that require fast infra. e. ----- Original Message -----