Hey, We have use case where we would like to load some data on a GCE persistent disk before running the main container. We'd like to run this as a separate step since the resource requirements differ significantly between the 'recovery' and normal operation. InitContainers allow us to ask for more resources during the recovery than the normal operation. However, this would require us to have bigger nodes running so the recovery will have sufficient resources. The difference in resource usage is on the scale of 16 CPUs during recovery to 2 CPUs during normal operation. Ideally, we'd like to have an autoscaling High CPU node pool, that will scale down to zero if there are no components in the recovery phase. Is this somehow possible to implement? We are running into a wall as taints and affinities are per Pod rather than per container and it seems that the initContainers are always running where the normal containers will be.
Can we somehow run the recovery phase on a high CPU node pool and run the main containers on the normal node pools? Regards, Sami Hiltunen -- You received this message because you are subscribed to the Google Groups "Kubernetes user discussion and Q&A" group. To unsubscribe from this group and stop receiving emails from it, send an email to kubernetes-users+unsubscr...@googlegroups.com. To post to this group, send email to kubernetes-users@googlegroups.com. Visit this group at https://groups.google.com/group/kubernetes-users. For more options, visit https://groups.google.com/d/optout.
