Well, the error message is telling you all you need to know: the scheduler cannot find placement for more than two Pods, because you only have two hosts, and you are using the default “hard”anti-affinity configuration. In that mode, the scheduler is not allowed to place two memcached Pods on the same node. Since you have fewer nodes (2) than the standard replica count (3), it can’t proceed.
Either use the “soft” mode, increase the size of your cluster, or decrease the required number of memcached Pods if you want this error to go away.