Details
-
Feature Request
-
Resolution: Done
-
Major
-
None
-
None
-
0
-
0%
Description
While JGRP-1570 only changes the default for desired_avg_gossip and max_bytes (not scaling them with the cluster size), this issue is about a new stable protocol, STABLE2, which works well in large clusters without generating too much traffic.
It works as follows:
- A dedicated member (the coord) periodically multicast a STABLE message
- Everyone replies with a (unicast) STABLE message and when the coord has received STABLE replies from everyone, it multicasts a STABILITY message
- This would only require a multicast from the coord to everyone, establishing TCP connections from the coord to everyone (usually already exists because of the VIEW-CHANGE multicast), but everyone would reuse the same TCP connection to send the reply
Another solution could be that every member sends its STABLE message to the current coordinator (triggered by time or size), and when the coordinator has received STABLE messages from all members, it multicasts a STABILITY message
Attachments
Issue Links
- is cloned by
-
JGRP-1570 STABLE: desired_avg_gossip leads to long intervals between reception of STABILITY messages in large clusters
- Resolved