[openstack-dev] [Gantt] [GSoC] Clarifications about Gantt opportunities for GSoC

Sylvain Bauza sylvain.bauza at gmail.com
Sun Mar 9 22:37:07 UTC 2014


I previously already sent an email but I'm recreating it in its own thread.
I just discovered that Gantt has been proposed as part of the ideas for
GSoC [1]
While I truly understand the mutual opportunities for both students and
Openstack, I'm thinking that Gantt is too young and yet to be stabiliized
before being ready for accepting applicants.

During the last weekly meetings we had, a strategy for deploying Gantt on
its roots has been defined and corresponding as follows :
 1. split out the dependencies from Nova by not proxying VM creation thru
the scheduler, but rather let the conductor place a call to the scheduler
and then spawn the VMs
 2. provide a scheduler library for resource tracking
 3. forklift Nova scheduler code into Gantt

As Nova is currently under Feature Freeze until Juno summit, and as most of
the design discussions will happen at the Summit, Gantt will not be
targeted to be delivered early in June, and will probably not be ready
until July/August.

Regarding the GSoC timeline (coding beginning on mid-May with a mid-time
evaluation on end-june)  [2], I consider Gantt as not matching with the
necessary requirements.

Following that, I ask GSoC proposed mentors to think about it and
reconsider Gantt as not an idea proposal for GSoC. That said, Nova
scheduler improvements (scalable scheduler, or others) can be considered as
good proposals but that's Nova, not Gantt responsability.

Feel free to comment,

[1] https://wiki.openstack.org/wiki/GSoC2014#Common_Scheduler_.28Gantt.29
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20140309/8221d2ec/attachment.html>

More information about the OpenStack-dev mailing list