← Back to team overview

fuel-dev team mailing list archive

Re: Fuel master node containerization update

 

Thanks Matt & team - it is great progress! Looking forward to see fully
automated master node install on Monday.

PS. Let's move conversations to openstack-dev, and keep only old threads
here.
On Apr 19, 2014 11:03 AM, "Matthew Mosesohn" <mmosesohn@xxxxxxxxxxxx> wrote:

> Hi all,
>
> We're quite close to our feature freeze (extended to 4/20) and I want
> to give a brief update.
>
> I have an environment locally deployed which can deploy OpenStack
> (CentOS, Havana) and run OSTF. There were a number of issues with
> missing explicit dependencies for astute and mcollective, but I've
> resolved them all either through package updates or puppet patches.
>
> The remaining issues to be met before feature freeze are these:
> 1 - finish polishing the bootstrap_admin_node.sh script. It extracts
> the lrz archive of docker images and loads them in currently, but
> doesn't deploy the containers in order. Deployment by hand by simply
> starting each one with dockerctl works just fine. I will spend my time
> this weekend on that task.
> 2 - Sergey O (OSCI) added openssh-clients as a dependency for astute
> RPM, but it's not in master repo yet.
> 3 - Tatyana L will update OSTF to provide a config file option for
> configuring parameters. If she can't deliver, we'll have to work extra
> variable passing into nailgun::supervisord puppet class to get OSTF
> all the required parameters for starting.
>
> The critical, but not blocker items for the release are:
> 1 - rework diagnostic snapshot so that it does not require nailgun to
> be on the same host as mcollective + shotgun
> 2 - finish development of Fuel ISO makefile that builds all containers
> from scratch (We can do it asynchronously for now and just download
> packaged containers)
> 3 - tuning of docker init script so that it kills all containers
> before it shuts down. It will prevent zombie process issues.
>
> Nice to haves if we can get them in time for the release:
> 1 - Re-optimize LRZ compression and decompression times (12 minutes to
> uncompress and docker load inside a VM is too long)
> 2 - Package dockerctl into an RPM
>
> I'd like to thank Alexander Didenko, Vladimir Sharshov, Vladimir
> Kuklin, Evgeniy Li, Andrey Danin, and our QA team for all their
> efforts this past week in getting all the code ready and helping to
> patch up bugs during the crunch.
>
> Best Regards,
> Matthew Moseoshn
>
> --
> Mailing list: https://launchpad.net/~fuel-dev
> Post to     : fuel-dev@xxxxxxxxxxxxxxxxxxx
> Unsubscribe : https://launchpad.net/~fuel-dev
> More help   : https://help.launchpad.net/ListHelp
>

References