In order to move PAWS to a magnum cluster the k8s parts will need to be replicated in parallel to the existing k8s parts. The existing cluster has 6 cpu and 12G ram for the control nodes, along with 44 cpu and 88G ram for the worker nodes. Ingress nodes will not be replicated. Additionally new nodes will use attached volume space rather than the disk space that would have come with older VMs, which is currently set to 100G for each worker node (there would be 11 to replicate this. We likely could make due with rather less, say 30G per worker.
The sum of which is to add 50 cpu and 100G ram to PAWS along with about 600G storage space.
This could be reduced a fair amount (And increase the amount that is returned when done) by running the work in T326260. As the cpu need would be cut down by 20 and then there would be 3 systems instead of 11 so existing storage should be enough. It would increase uncertainty somewhat, but seems reasonable. Let's go with that.
So this is a request to temporarily increase PAWS quota by 30 cpu and 108G ram