I'm testing a following configuration on latest percona xtra db cluster with the following settings:
Zone A (sql nodes A1 through A3 are 10.10-12) and following setting on cluster address: wsrep_cluster_address=gcom://10.0.0.10,10.0.0.11,10.0.0.12
Zone B (sql node B1 is 10.0.1.10) and following setting on cluster address: wsrep_cluster_address=gcom://10.0.0.10,10.0.0.11,10.0.0.12
So clearly B1 isn't participating in the cluster (intended behavior), but does receive the updates from the cluster. I never want B1 to be part in quorum determination and just sit there and be available when Zone gets nuked (manual switch at that point in time).
My B1 node now gets error of WSREP has not yet prepared node for application use but really I don't care about it (or should I ?) while its not in use.
I've read variety of articles on correct number of nodes and its pros/cons (ie one here https://www.percona.com/blog/2015/06/23/percona-xtradb-cluster-pxc-how-many-nodes-do-you-need/) , but what is a potential problem with this setup ? I also don't want to have another set of nodes in Zone B, but rather just one.
https://stackoverflow.com/questions/66881106/percona-db-cluster-replication-to-a-node-outside-the-wsrep-cluster-address-gcomm March 31, 2021 at 11:07AM
没有评论:
发表评论