Cluster access for Firmament/Poseidon about cluster HOT 10 CLOSED

+1

from cluster.

On behalf of @cncf/intelclusterteam - is this request valid?

If so, we have two questions:

1. Unfortunately, the maximum amount of nodes available at the moment is ~400 and it probably won't change a lot (i.e. it won't reach 500) until expansion which should take place in Q1'17. Are 400 nodes enough for the second part of your tests?
2. I understand this request as follows - you would need 1 week with 100 nodes and then another week with 400 nodes. In total - two weeks of cluster usage, week after week. Is that correct?

from cluster.

[ adding @ICGog, who is doing Firmament experimentation and drives Poseidon development ]

Apologies for the late response -- other projects intervened! The request is still valid (albeit not urgent), so would it be okay to reopen it?

As for the questions:

1. 400 nodes are sufficient for our scale-up tests.
2. Yes, that would be ideal, but we can be flexible.

from cluster.

Hi @ms705 , no problem at all - we just thought that it's no longer valid.

We will put this request in the queue after the maintenance window, once we know how many nodes are available afterwards then we will update you on the exact dates.

from cluster.

Hi @ms705, we are wondering if you would like to start working after Red Hat leaves the cluster at the end of March or would you rather have the nodes allocated later?

from cluster.

/cc @ICGog

Thanks! Ionel and I both have major deadlines coming up in late April, so having a window a bit later in the year would actually suit us better (if possible).

from cluster.

Hi @ms705, how does your schedule look like in May and June? What would be the best time for you to start your research on CNCF cluster?

Currently we have ~230-280 available nodes (this may change a bit), so from our perspective next few weeks are definitely a good time for arranging the nodes for you.

Let us know! :-)

from cluster.

Thanks a lot for the offer! The next few weeks are a bit too early for us. We've ported our Kubernetes integration to golang and we need another two months to stabilise it and to develop the workloads we want to run. It would be ideal if we could run the experiments at the end of July or in August. Would that work for you?

from cluster.

@ICGog Please plan to complete all work on the cluster by mid-July, as we may lose availability beyond that.

from cluster.

we have a new cluster up, announced here: https://www.cncf.io/community/infrastructure-lab/

we encourage folks to reapply if they are interested in access

from cluster.