3 ms·
What this article fails to talk about are primarily the failure scenarios of the queues under partitions. Does the queue become write-unavailable, read-unavaila
by sargun 12y ago
What this article fails to talk about are primarily the failure scenarios of the queues under partitions. Does the queue become write-unavailable, read-unavailable, inconsistent, or otherwise during network partition? Can the queue remain partially available during network partition?
These are the situations, in a scaling, production system that really matter.
- adamw1pl 12y agoThese are very good questions, however for a different post. Here I was mainly interested in the throughput. Jepsen post provide some insights (into Mongo and Rabbit). With Hornet you can easily end up with two primaries. SQS is untestable in that respect :).