• RadonDB

  • (已解决)v2.2.1 升级 v2.3.0 init-sidecar 报错

sidecar 日志

panic: invalid mysql version, currently we only support 5.7 or 8.0

goroutine 1 [running]:
github.com/radondb/radondb-mysql-kubernetes/sidecar.NewInitConfig(0x15cb920)
	/workspace/sidecar/config.go:140 +0x10eb
main.main()
	/workspace/cmd/sidecar/main.go:91 +0x116

MysqlCluster

apiVersion: mysql.radondb.com/v1alpha1
kind: MysqlCluster
metadata:
  name: openldap
spec:
  replicas: 3
  mysqlVersion: "5.7"
  mysqlOpts:
    mysqlConf:
      expire_logs_days: "7"
      innodb_buffer_pool_instances: "4"
    resources:
      requests:
        cpu: 100m
        memory: 256Mi
      limits:
        cpu: 1
        memory: 2Gi

  metricsOpts:
    enabled: true

  podPolicy:
    slowLogTail: true
    auditLogTail: true

  persistence:
    enabled: true
    size: 10Gi

Helm

helm upgrade -i radondb radondb/mysql-operator --version 2.3.0

升级后发现 mysqlclusters 中镜像还是 v2.2.1 的，手动替换后，错误依旧。

你好,旧版本升级需要分以下步骤:
1. 升级crd, helm upgrade只能升级operator镜像, helm无法更新crd的内容(工具目前为止都不支持)
kubectl apply https://github.com/radondb/radondb-mysql-kubernetes/tree/v2.3.0/config/crd/bases
链接中的三个crd
2. 执行helm upgrade -i radondb radondb/mysql-operator –version 2.3.0
3. 升级已有的集群:
下载 https://github.com/radondb/radondb-mysql-kubernetes/blob/v2.3.0/config/samples/mysql_v1alpha1_mysqlcluster_mysql8.yaml
修改为原集群对应的参数,镜像的tag 改为v2.3.0
4 检查集群pod 是否正常:
kubectl get po -Lrole,healthy , healthy需为yes

    9 个月 后
    jangrui 更改标题为「(已解决)v2.2.1 升级 v2.3.0 init-sidecar 报错」