云仓k8s版本部署和卸载步骤

一、部署gcdw步骤

1.     修改/opt/gcdw-k8s-9.8.0.4.1/helm/helm_gcdw/chart的配置文件values.yaml，如下：

灰色阴影部分是需要修改的部分，其他可使用默认配置：

# Default values for gcdw.

# This is a YAML-formatted file.

# Declare variables to be passed into your templates.

global:

  # images setting

  image:

    registry: "harbor.gbase.cn"               =========>仓库地址

    pullPolicy: Always

    imagePullSecrets: ""                     =========>仓库密钥，公有仓库不用填

  # Tenant mode for gcdw, true: multiple-tenant; false: single-tenant

  multiTenantMode:

    enabled: false                       ======>是否开启多租户模式,不开启时为单租户模式，开启后为多租户模式

    imageRegistry:

      server: ""                          ======>仓库地址

      username: ""                       ======>访问仓库的用户名

      password: ""                        ======>访问仓库的密码

  # fdb setting

  metadata:

    fdb_cluster: "WQSxlF05:Ujlj3egX@10.10.7.75:4500"       ======>fdb服务

  # persistant storage setting for gcware

  storage:

    enabled: false

    storageClass: ""

    storageSize: ""

  # logging system setting

  logging:

    enabled: false

    filebeatImage: ""

    filebeatOutput:

      output.elasticsearch:

        hosts:

        - ""

        username: ""

        password: ""

        index: 'gcdw-${MY_POD_NAMESPACE}-${MY_POD_NAME}-log-%{+yyyy.MM.dd}'

    logrotateImage: "gcdw/gcdw-logrotate"

  # node affinity setting

  nodeAffinitySchedule:

    cloudServiceAffinity:

      key:  cloudService

      value: gcdw

    coordiantorServiceAffinity:

      key: coordinatorService

      value: gcdw

    warehouseServiceAffinity:

      key: warehouseService

      value: gcdw

 # kerberos setting

  kerberos:

    enabled: true

  # # Namespace label setting

  # namespaceLabel: "Organization=gcdw"

# gcdw operator values-------------

operator:

  enabled: true

  image:

    repository: gcdw/gcdw-operator

    tag: 9.8.0.4.1

  replicas: 1

  leaderElect: true

  resources:

    requests:

      cpu: 1

      memory: 1Gi

    limits:

      cpu: 2

      memory: 2Gi

# gcdw server values-------------------

server:

  image:

    repository: gcdw/gcdw-server

    tag: 9.8.0.4.1

  gcware_replicas: 3                 =====>初始部署的gcware个数

  coordinator_replicas: 2             ======>初始部署的coor个数

  gcware_resource:                 =====>以下是gcware、coor、warehosue的大小设置，根据具体需要配置

      request:

        cpu: "1"

        mem: "2Gi"

      limit:

        cpu: "4"

        mem: 4Gi

  coordinator_resource:

      request:

        cpu: "2"

        mem: 4Gi

      limit:

        cpu: "4"

        mem: 16Gi

  warehouse_resource:

      request:

        cpu: "2"

        mem: 4Gi

      limit:

        cpu: "4"

        mem: 16Gi

# gcdw saas values--------------------

saas:

  enabled: true

  # set time zone env for saas-backend

  TZ: "Asia/Shanghai"

  image:

    frontend:

      repository: gcdw/clound-database-frontend

      tag: 9.8.0.4.1

    backend:

      repository: gcdw/clound-database-backend

      tag: 9.8.0.4.1

    redis:

      repository: gcdw/redis

      tag: latest

    vsftpd:

      repository: gcdw/vsftpd

      tag: latest

  replicas:

    frontend: 1

    backend:  1

    redis: 1

    vsftpd: 1

  webServicePort: 32144      =====>前端段口号

2.     创建命名空间

kubectl create ns gcdwtest2

3.     在/opt/gcdw-k8s-9.8.0.4.1/helm/helm_gcdw目录下进行安装

helm install gcdw chart/ -n gcdwtest2

4.     查看命名空间，等待基础服务组件已安装成功后注册租户

[root@liurui-node-2 helm_gcdw]# kubectl get all -n gcdwtest2

NAME                                 READY   STATUS    RESTARTS   AGE

pod/gcdw-ftp-0                       1/1     Running   0          26s

pod/gcdw-operator-5d6c5857bb-xlx46   1/1     Running   0          28s

pod/gcdw-redis-0                     1/1     Running   0          26s

pod/gcdw-saas-backend-0              1/1     Running   0          26s

pod/gcdw-saas-frontend-0             1/1     Running   0          26s

注册租户(前端或后端注册均可以）

前端入口:10.10.7.75:32145（配置文件中设置的端口号）

后端注册的方式：

①进入pod/gcdw-saas-backend-0

kubectl exec -it gcdw-saas-backend-0 -n gcdwtest -- bash

②注册租户

./account --create -A testaccount -U root -P 111111 -C gcdw -M 18812345678 -E langzhishuang@gbase.cn -N gcdw --S3_accessKey=BW2L6QN5JV60R160QGTV --S3_secretKey=fDY4NtDGh5Mf8zVKD2+w9yeZGZTsDngkExsZihu1 --S3_endPoint=10.10.7.75:9000 --S3_region=minio --S3_bucket=song --comment=test

5. 查看租户是否注册成功

[root@liurui-node-2 helm_gcdw]# kubectl get all -o wide  -n test

NAME                                 READY   STATUS    RESTARTS   AGE

pod/coordinator-1-0                  1/1     Running   0          48m

pod/coordinator-2-0                  1/1     Running   0          48m

pod/gcdw-ftp-0                       1/1     Running   0          4h23m

pod/gcdw-operator-5d6c5857bb-sr5gf   1/1     Running   0          52m

pod/gcdw-redis-0                     1/1     Running   0          4h23m

pod/gcdw-saas-backend-0              1/1     Running   0          4h23m

pod/gcdw-saas-frontend-0             1/1     Running   0          4h23m

pod/gcware-0                         1/1     Running   0          48m

pod/gcware-1                         1/1     Running   0          48m

pod/gcware-2                         1/1     Running   0          48m

进入coordinator-1-0并使用集群：

[root@liurui-node-2 ~]# kubectl exec -it coordinator-1-0 -n test -- bash

[gbase@coordinator-1-0 /]$ gccli -uroot -p111111

GBase client 9.8.0.4.17ff7898. Copyright (c) 2004-2023, GBase.  All Rights Reserved.

gbase> create warehouse wh1;

Query OK, 0 rows affected (Elapsed: 00:00:00.26)

gbase> alter warehouse wh1 resume;

Query OK, 0 rows affected (Elapsed: 00:00:07.78)

gbase> alter user root default_warehouse='wh1';

Query OK, 0 rows affected (Elapsed: 00:00:00.09)

gbase> \q

Bye

租户注册成功且集群能正常使用。

6. 多租户模式需开启values.yaml配置文件中的参数；

multiTenantMode:

    enabled: true

该模式下可以注册多个租户，但租户名与命名空间名称不能相同，单租户则要求命名空间要与注册租户名字一致。

注册多个租户(这里注册syl1和syl2作为示例)

(1)注册租户syl1

注册租户(前端或后端注册均可以，前端入口:10.10.55.74:32168，这里展示后端注册的方式)

①进入pod/gcdw-saas-backend-0

kubectl exec -it gcdw-saas-backend-0 -n test -- bash

②注册租户

./account --create -A syl1 -U root -P 111111 -C gcdw -M 13042299268 -E songyanling@gbase.cn -N gcdw --S3_accessKey=BW2L6QN5JV60R160QGTV --S3_secretKey=fDY4NtDGh5Mf8zVKD2+w9yeZGZTsDngkExsZihu1 --S3_endPoint=10.10.55.36:9000 --S3_region=minio --S3_bucket=song1 --comment=test

(2) 注册租户syl2

①进入pod/gcdw-saas-backend-0

kubectl exec -it gcdw-saas-backend-0 -n gcdwtest2 -- bash

②注册租户

./account --create -A account1 -U root -P 111111 -C gcdwtest2 -M 13042299268 -E langzhishuang@gbase.cn -N gcdwtest2 --S3_accessKey=BW2L6QN5JV60R160QGTV --S3_secretKey=fDY4NtDGh5Mf8zVKD2+w9yeZGZTsDngkExsZihu1 --S3_endPoint=10.10.7.75:9000 --S3_region=minio --S3_bucket=gbase --comment=gcdwtest2

7. 查看syl1和syl2租户是否注册成功

(1)查看syl1租户是否注册成功及集群是否可用

[root@liurui-node-2 helm_gcdw]# kubectl get all -n syl1

NAME                  READY   STATUS    RESTARTS   AGE

pod/coordinator-1-0   1/1     Running   0          26m

pod/coordinator-2-0   1/1     Running   0          26m

pod/gcware-0          1/1     Running   0          26m

pod/gcware-1          1/1     Running   0          26m

pod/gcware-2          1/1     Running   0          26m

进入coordinator-1-0并使用集群：

[root@liurui-node-2 ~]# kubectl exec -it coordinator-1-0 -n syl1 -- bash

[gbase@coordinator-1-0 /]$ gccli -uroot -p111111

GBase client 9.8.0.4.17ff7898. Copyright (c) 2004-2023, GBase.  All Rights Reserved.

gbase> create warehouse wh1;

Query OK, 0 rows affected (Elapsed: 00:00:00.26)

gbase> alter warehouse wh1 resume;

Query OK, 0 rows affected (Elapsed: 00:00:07.78)

gbase> alter user root default_warehouse='wh1';

Query OK, 0 rows affected (Elapsed: 00:00:00.09)

gbase> \q

Bye

[gbase@coordinator-1-0 /]$ gccli -uroot -p111111

GBase client 9.8.0.4.17ff7898. Copyright (c) 2004-2023, GBase.  All Rights Reserved.

gbase> create database test;

Query OK, 1 row affected (Elapsed: 00:00:00.28)

gbase> use test;

Query OK, 0 rows affected (Elapsed: 00:00:00.01)

gbase>  CREATE TABLE lineitem (

    ->   "l_orderkey" bigint(20) DEFAULT NULL,

    ->   "l_partkey" bigint(20) DEFAULT NULL,

    ->  "l_suppkey" bigint(20) DEFAULT NULL,

    ->   "l_linenumber" bigint(20) DEFAULT NULL,

    ->    "l_quantity" decimal(15,2) DEFAULT NULL,

    ->   "l_extendedprice" decimal(15,2) DEFAULT NULL,

    ->   "l_discount" decimal(15,2) DEFAULT NULL,

    ->     "l_tax" decimal(15,2) DEFAULT NULL,

    ->    "l_returnflag" char(1) DEFAULT NULL,

    ->    "l_linestatus" char(1) DEFAULT NULL,

    ->   "l_shipdate" date DEFAULT NULL,

    ->    "l_commitdate" date DEFAULT NULL,

    ->    "l_receiptdate" date DEFAULT NULL,

    ->   "l_shipinstruct" char(25) DEFAULT NULL,

    ->     "l_shipmode" char(10) DEFAULT NULL,

    ->   "l_comment" varchar(50) DEFAULT NULL

    -> ) ENGINE=EXPRESS  DISTRIBUTED BY('l_orderkey');

Query OK, 0 rows affected (Elapsed: 00:00:00.10)

gbase> load data infile 'sftp://gbase:gbase@10.10.14.88/opt/lineitem.tbl' into table test.lineitem fields terminated by '|' ;

Query OK, 6001215 rows affected (Elapsed: 00:00:14.54)

Task 3 finished, Loaded 6001215 records, Skipped 0 records

syl1租户注册成功且集群能正常使用。

(2)查看syl2租户是否注册成功及集群是否可用

[root@liurui-node-2 helm_gcdw]# kubectl get all -n syl2

NAME                  READY   STATUS    RESTARTS   AGE

pod/coordinator-1-0   1/1     Running   0          77s

pod/coordinator-2-0   1/1     Running   0          76s

pod/gcware-0          1/1     Running   0          77s

pod/gcware-1          1/1     Running   0          77s

pod/gcware-2          1/1     Running   0          77s

进入coordinator-1-0并使用集群：

[root@liurui-node-2 ~]# kubectl exec -it coordinator-1-0 -n syl2 -- bash

[gbase@coordinator-1-0 /]$ gccli -uroot -p111111

GBase client 9.8.0.4.17ff7898. Copyright (c) 2004-2023, GBase.  All Rights Reserved.

gbase> create warehouse wh1;

Query OK, 0 rows affected (Elapsed: 00:00:00.26)

gbase> alter warehouse wh1 resume;

Query OK, 0 rows affected (Elapsed: 00:00:07.78)

gbase> alter user root default_warehouse='wh1';

Query OK, 0 rows affected (Elapsed: 00:00:00.09)

gbase> \q

Bye

[gbase@coordinator-1-0 /]$ gccli -uroot -p111111

GBase client 9.8.0.4.17ff7898. Copyright (c) 2004-2023, GBase.  All Rights Reserved.

gbase> create database test;

Query OK, 1 row affected (Elapsed: 00:00:00.28)

gbase> use test;

Query OK, 0 rows affected (Elapsed: 00:00:00.01)

gbase>  CREATE TABLE lineitem (

    ->   "l_orderkey" bigint(20) DEFAULT NULL,

    ->   "l_partkey" bigint(20) DEFAULT NULL,

    ->  "l_suppkey" bigint(20) DEFAULT NULL,

    ->   "l_linenumber" bigint(20) DEFAULT NULL,

    ->    "l_quantity" decimal(15,2) DEFAULT NULL,

    ->   "l_extendedprice" decimal(15,2) DEFAULT NULL,

    ->   "l_discount" decimal(15,2) DEFAULT NULL,

    ->     "l_tax" decimal(15,2) DEFAULT NULL,

    ->    "l_returnflag" char(1) DEFAULT NULL,

    ->    "l_linestatus" char(1) DEFAULT NULL,

    ->   "l_shipdate" date DEFAULT NULL,

    ->    "l_commitdate" date DEFAULT NULL,

    ->    "l_receiptdate" date DEFAULT NULL,

    ->   "l_shipinstruct" char(25) DEFAULT NULL,

    ->     "l_shipmode" char(10) DEFAULT NULL,

    ->   "l_comment" varchar(50) DEFAULT NULL

    -> ) ENGINE=EXPRESS  DISTRIBUTED BY('l_orderkey');

Query OK, 0 rows affected (Elapsed: 00:00:00.10)

gbase> load data infile 'sftp://gbase:gbase@10.10.14.88/opt/lineitem.tbl' into table test.lineitem fields terminated by '|' ;

Query OK, 6001215 rows affected (Elapsed: 00:00:14.54)

Task 3 finished, Loaded 6001215 records, Skipped 0 records

syl2租户注册成功且集群能正常使用。

二、卸载gcdw步骤

1.删除fdb中租户信息

kubectl exec -it 进入到coordinator的pod上执行 accout drop -A租户名

account --drop -A syl1

如果执行account drop -A 租户名命令报错时，先执行：

export GBASE_NOT_CREATE_ACCOUNT_RESOURCE=1

2.执行卸载命令：helm uninstall gcdw -n

3.删除namespace，也可以不删，再次部署时可以继续使用。

删除命名空间：kubectl delete ns test