Scylla Manager using Helm and configure S3 storage

Kubernetes Operator
, , ,
1

Hi everyone,

I’m trying to deploy Scylla Manager using Helm and configure S3 storage with Wasabi. Below is my values.manager.yaml:

image:
  tag: 3.3.0

resources:
  limits:
    cpu: 200m
    memory: 256Mi
  requests:
    cpu: 200m
    memory: 32Mi

controllerResources:
  limits:
    cpu: 200m
    memory: 128Mi
  requests:
    cpu: 200m
    memory: 32Mi

scylla:
  developerMode: true
  scyllaImage:
    tag: 5.4.3
  agentImage:
    tag: 3.3.0
  datacenter: manager-dc
  racks:
    - name: manager-rack
      members: 1
      storage:
        capacity: 5Gi
        storageClassName: longhorn
      resources:
        limits:
          cpu: 1
          memory: 200Mi
        requests:
          cpu: 1
          memory: 200Mi
      scyllaAgentConfig: scylla-agent-config

I also created a scylla-agent-config ConfigMap with the following content:

yaml

Copy code

s3:
  access_key_id: <key>
  secret_access_key: <secret>
  provider: "Minio"
  endpoint: "https://s3.ap-southeast-1.wasabisys.com"

When I attempt to take a backup using:

bash

Copy code

sctool backup -c scylla/scylla -L 's3:astring-backup' --retention 7 -i 24h

I get this error:

yaml

Copy code

Flag --interval has been deprecated, use cron instead
Error: create backup target: location is not accessible
10.43.119.197: giving up after 2 attempts: after 30s: context deadline exceeded
Trace ID: V4nAJ30MS5SNxrODuV9TXA (grep in scylla-manager logs)

What configurations or steps am I missing to successfully configure Scylla Manager with Wasabi S3 storage so my backups work?

2

Overall scyllaAgentConfig needs to reference a Secret, not a ConfigMap, and needs to be set up for the target ScyllaDB cluster, not the one used by Scylla Manager. With just these snippets it’s hard to say if these are all the changes required to wire it correctly, so please create an issue with must-gather archive attached.

1 Like