I agree it’s more confortable for the CI but it’s more complicated for the provider. As CI have needs, we should think about the simplest solution on the provider side if we want more ressources at the infrastructure donors.
I undertand that multi node jobs are difficult to manage on spot instances. Do you have something to tag jobs and manage different kind of ressources? You cloud have « spot jobs » which could be killed and restart easily at any time, and you could have classic jobs which require to terminate normally. Does it make sense? — Jean-Daniel Bonnetot http://www.ovh.com @pilgrimstack > Le 17 déc. 2015 à 23:41, Jeremy Stanley <[email protected]> a écrit : > > On 2015-12-17 14:27:45 -0800 (-0800), James E. Blair wrote: > [...] >> While I think this approach could work, it doesn't feel like a >> natural fit for me. Our tests and therefore instances are actually >> rather short-lived. Rather than using a system where, at 57 >> minutes into a 60 minute job, the instance is deleted and we need >> to restart the work, it would be more ideal to simply let that job >> complete, let the instance be naturally deleted, and not replace >> it. In other words, a smoother system for us that may achieve >> similar results more efficiently would be one that simply adjusted >> our quota dynamically. > > I completely agree that given a choice between the two, a graceful > adjustment (perhaps with guarantees on the Nova side so the provider > can confirm deletion of oldest instances at some time window after > the quota drop) would be preferable over something that just deletes > random instances and hides the effective lower quota we end up at. > -- > Jeremy Stanley _______________________________________________ OpenStack-Infra mailing list [email protected] http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-infra
