Previously, we enabled maintenance mode for HAProxy on the Postgres servers of the cluster, which could cause errors in configurations that use dedicated load balancers. Now, we delegate these tasks directly to the HAProxy load balancer servers.
Fixed:
TASK [upgrade : Stop confd service] ********************************************************************************************************************************************************************************************************************
fatal: [***.**.**.***]: FAILED! => {"changed": false, "msg": "Could not find the requested service confd: host"}
fatal: [***.**.**.***]: FAILED! => {"changed": false, "msg": "Could not find the requested service confd: host"}
Additionally, delete the patroni.dynamic.json file before preparing the configuration for the new cluster.
Issue: https://github.com/vitabaks/postgresql_cluster/issues/666
Previously, we enabled maintenance mode for HAProxy on the Postgres servers of the cluster, which could cause errors in configurations that use dedicated load balancers. Now, we delegate these tasks directly to the HAProxy load balancer servers.
Fixed:
Additionally, delete the
patroni.dynamic.json
file before preparing the configuration for the new cluster.