About migrate to causal cluster for HA clusters

34 views
Skip to first unread message

Liping Huang

unread,
Jan 8, 2018, 4:04:05 AM1/8/18
to Neo4j
Hi there,

If I already have a neo4j HA clusters, let's say neo4j_1(Master), neo4j_2, neo4j_3,
1. then I shutdown the master server neo4j_1 to import new data,
2. config the neo4j_1 as the "CORE" for Causal Cluster mode as the seed server
3. then I start neo4j_1 to add to the existing HA clusters

is is OK? will neo4j_1 seed neo4j_2 and neo4j_3 without restart/reconfig neo4j_2 and neo4j_3? or oppositely neo4j_1 will take the data from neo4j_2 or neo4j_3?

Regards.

Michael Hunger

unread,
Jan 8, 2018, 4:25:35 AM1/8/18
to ne...@googlegroups.com, Mark Needham
You will have to shut down and configure all 3 as core instances (minimum 3).

Also, make sure the ports for Causal Cluster are properly configured on each machine.

And start the cluster again.

--
You received this message because you are subscribed to the Google Groups "Neo4j" group.
To unsubscribe from this group and stop receiving emails from it, send an email to neo4j+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Liping Huang

unread,
Jan 9, 2018, 2:39:32 AM1/9/18
to Neo4j
Thanks Michael,

shutdown all of them is not a choose for me, in a HA cluster, how to re-seed the cluster data in online mode? take above neo4j instances as an example:
1. Delete all data in neo4j_1 ( the master instance )
2. reload data into neo4j_1

 then how the synchronization works? 
1. When delete all the data in master instance
    1) all the data stored in slaves will be deleted too?
    2) delete the master data willl not impact the slaves?
2. the reloaded data in master server will sync to slaves by default.

Thanks in advance.


在 2018年1月8日星期一 UTC+8下午5:25:35,Michael Hunger写道:
To unsubscribe from this group and stop receiving emails from it, send an email to neo4j+un...@googlegroups.com.

Michael Hunger

unread,
Jan 9, 2018, 6:17:40 AM1/9/18
to ne...@googlegroups.com, Mark Needham
Yes if you delete data in the master it will be deleted in the slaves too (it's replication)

You can seed a core instance / read-replica with the lastest master / leader state 
if you don't then the new cluster members without any data would pull their database from the cluster.

I don't think you can do migration from HA to CC without downtime.

Mark might have some more pointers?

Michael

To unsubscribe from this group and stop receiving emails from it, send an email to neo4j+unsubscribe@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages