Donno about hiccups, but the memcached servers are not replicas: each has half the segments

But this is really two different questions

What you really want to know is how will peers behave when a memcached becomes unreachable

I know everything will still work but I’m unsure if there will be extra blocking timeouts added to peer work or not.

there will not be blocking timeouts

the memcache response timeout is very short

if it doesn’t return within that very short window the peer will go to storage instead

Thanks a lot for the explanations! helped in my understanding!