When deployed with clickhouse operator, is it possible to run clickhouse backup on only one replica? about clickhouse-backup HOT 2 CLOSED

by default, clickhouse-backup requires memory only when upload/download data usually memory fooprint is only 30-40mb without these operations

issue is not related to clickhouse-backup itself

you can specify a different podTemplate for each replica in shard

something like that

spec:
  templates:
  - name: pod-with-backup
    spec:
      containers:
      - name: clickhouse-backup
        image: altinity/clickhouse-backup:latest
      - name: clickhouse
        image: clickhouse/clickhouse-server:latest  
  - name: pod-clickouse-only
    spec:
      containers:
      - name: clickhouse
        image: clickhouse/clickhouse-server:latest  
  configuration:
    layout:
      shardsCount: 2 
      replicas:
      - templates:
          podTemplate: pod-with-backup
      - templates:
          podTemplate: pod-clickhouse-only

Moreover, it will more complex command sequence to restore data in this cluster in this case
need to add environment variable RESTORE_SCHEMA_ON_CLUSTER="{cluster}" in clickhouse-backup container spec

from clickhouse-backup.

@Slach thanks for example with with replicas that looks like exactly what we need here.

clickhouse-backup requires memory only when upload/download data usually memory fooprint is only 30-40mb without these operations

We are planning on using blob storage with relatively large tables, so our ch-backup sidecar will need non-trivial resources.

Moreover, it will more complex command sequence to restore data in this cluster in this case
need to add environment variable RESTORE_SCHEMA_ON_CLUSTER="{cluster}" in clickhouse-backup container spec

We were planning on manually recreating schema on replicas w/o the backup sidecar, but this may be a better option.

from clickhouse-backup.