When there are some problems in the RisingWave cluster, such as node restart and network abnormalities, the cluster will enter recovery processing to ensure data consistency.
alter system set pause_on_next_bootstrap to true;
risectl
, refer to: Cluster scaling.risectl meta resume
.waiting for new workers to join, elapsed xxxs
.-parallelism
, or temporarily launch some compute nodes to meet the requirements of parallelism.risectl meta unregister-workers --workers <worker_id or worker_host, ...>
.-parallelism
and start the compute nodes.auto scaling
feature that is on the way.
connection refused
or error trying to connect: dns error: failed to lookup address information: Name or service not known
.
Solution:
Please check the network configuration of the deployment environment and whether the operators of k8s are working properly.