You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The way how a cluster is being rebalanced by adding new/empty node is pretty straightforward and clear.
I wanted to ask what is the best flow for rebalancing an existing cluster (replication=1) by removing a node?
I tried switch/setup replication=2 (on both carbon-c-relay and buckyd) on running cluster but bucky inconsistent does not show any inconsistency.
The text was updated successfully, but these errors were encountered:
I never got replication > 1 support working in Bucky tools before the client I was working with went in different directions, unfortunately.
To move to a smaller Graphite cluster, I would usually build that smaller cluster and use rebalance to transition to it. But I think the same can be done to just remove a node from the existing cluster. (Watch out if using the Jump hash method as it really only supports removing the last node rather than a node in the middle.)
Allows you to specify additional buckyd daemons (graphite storage hosts) that are not in the new hashring and consider those metrics when rebalancing the hashring.
The way how a cluster is being rebalanced by adding new/empty node is pretty straightforward and clear.
I wanted to ask what is the best flow for rebalancing an existing cluster (
replication=1
) by removing a node?I tried switch/setup
replication=2
(on bothcarbon-c-relay
andbuckyd
) on running cluster butbucky inconsistent
does not show any inconsistency.The text was updated successfully, but these errors were encountered: