Couchbase HA issues

[sebarys]

Hello,

We are testing Couchbase cluster HA and having some questions:

• When using Automatic Failover it is possible to rebalance cluster automatically after node failover?
• If node will come back it is possible to add it to the cluster automatically and rebalance cluster?
• We expected that when master node of some vBucket will went down other node which have replica will be elected automatically, unfortunately we observed that when node went down we have downtime. It is by design or maybe our configuration is wrong (default configuration)?

Thank you for help and suggestions

[sebarys]

UP ^
Can anyone provide me some informations?

[househippo]

@sebarys,

@sebarys

When using Automatic Failover it is possible to rebalance cluster automatically after node failover?

you can script rebalance if needed.

@sebarys

If node will come back it is possible to add it to the cluster automatically and rebalance cluster?

Most people want to do a RCA of why a node was kicked out of a node before adding it back in

@sebarys

We expected that when master node of some vBucket will went down other node which have replica will be elected automatically, unfortunately we observed that when node went down we have downtime. It is by design or maybe our configuration is wrong (default configuration)?

What is your autofailure set to 30 seconds ? 120 Seconds? in CB 5.0 you can set it as low as 5 seconds, but please test as network false positives might cause issues.