3 ms·
I'm curious to know more about the garbage collection of stale peers. - AFAIK, etcd is built on RAFT, which relies on a 'joint majority' method of transitionin
by bkirwi 13y ago
I'm curious to know more about the garbage collection of stale peers.
- AFAIK, etcd is built on RAFT, which relies on a 'joint majority' method of transitioning cluster membership. Are there any issues forming agreement on what the new membership should be when it's unclear what nodes are still supposed to be part of the cluster?
- In the land of ZooKeeper, cluster configurations are typically very very stable, so tracking membership takes very little information. Is etcd targeted to more dynamic environments where the garbage generated by entering and leaving nodes is significant?
- philips 13y agoetcd isn't designed for significantly dynamic environments but we certainly want to make it easy for people to add and remove machines at runtime. We have basic support for runtime cluster configuration but are building out a more complete API. The current proposal that is being built can be found over here: http://thread.gmane.org/gmane.comp.distributed.etcd/168 http://thread.gmane.org/gmane.comp.distributed.etcd/168 As you point out there is no magic garbage collection of stale peers. You have to explicitly delete nodes that you don't expect to be participating any longer.