Open Stack

>
> I think in a four device single node single replica setup I'd probably just
> run request_node_count = 4 and call it a day.

I'll give this a shot right away.

But there are two questions that remain unanswered.

1. Why is there a discrepancy in the way writes vs reads are handled?
Isn't request_node_count honored during writes?
2. The same problem could happen anytime when # of nodes > 2 * replica
count, right? If there are 10 nodes and 3 replicas, could a write
happen to the 7th handoff node? If so, the read will fail because it
will only try 6 nodes.

I'll update this thread with the results of bumping request_node_count.