5.4. 测试 2：使用被动第三个站点的主节点故障切换

Expand

测试的主题	第三站点没有注册。即使第三个站点停机，故障转移也可以正常工作。
测试先决条件	DC1 上的 SAP HANA，DC2 正在运行并在 DC3 上停止。集群已启动并运行，且没有错误或警告。
测试步骤	使用 `pcs move` 命令移动 SAPHana 资源。
启动测试	执行集群命令：+ `pcs move resource SAPHana_MTR_00-clone`
监控测试	在第三个站点作为 `sidadm`:`% watch hdbnsutil -sr_stateConfiguration`运行在集群节点中以 root 用户身份运行：{ `watch pcs status`
预期结果	DC3 没有更改。
返回初始状态的方法	在新主上重新注册 DC3，并启动 SAP HANA。

专家摘要：

在第三个节点上停止 SAP HANA：
```
% HDB stop
```
```
% HDB stop
```
Copy to Clipboard Toggle word wrap

检查节点 3 上的集群状态和主节点：

pcs status --full
% hdbnsutil -sr_stateConfiguration | grep "primary masters"

# pcs status --full
% hdbnsutil -sr_stateConfiguration | grep "primary masters"

Copy to Clipboard

Toggle word wrap

在集群节点上启动故障切换：
```
pcs resource move <SAPHana-clone-resource>
```
```
# pcs resource move <SAPHana-clone-resource>
```
Copy to Clipboard Toggle word wrap

检查主要是否已移动，但在第三个节点上，主节点没有改变：

pcs resource  | grep Promoted
% hdbnsutil -sr_stateConfiguration | grep "primary masters"

# pcs resource  | grep Promoted
% hdbnsutil -sr_stateConfiguration | grep "primary masters"

Copy to Clipboard

Toggle word wrap

清理环境以再次运行测试：
```
psc resource clear <SAPHana-clone-ressource>
```
```
# psc resource clear <SAPHana-clone-ressource>
```
Copy to Clipboard Toggle word wrap

使用新主注册 node3 ：

% hdbnsutil -sr_register --remoteHost=node2 --remoteInstance=${TINSTANCE} --replicationMode=syncmem --name=DC3 --remoteName=DC2 --operationMode=logreplay --onlin

% hdbnsutil -sr_register --remoteHost=node2 --remoteInstance=${TINSTANCE} --replicationMode=syncmem --name=DC3 --remoteName=DC2 --operationMode=logreplay --onlin

Copy to Clipboard

Toggle word wrap

在 node3 上启动 SAP HANA：
```
% HDB start
```
```
% HDB start
```
Copy to Clipboard Toggle word wrap

输出详情的步骤

停止 node3 上的数据库：

sidadm+@node3% HDB stop
hdbdaemon will wait maximal 300 seconds for NewDB services finishing.
Stopping instance using: /usr/sap/RH2/SYS/exe/hdb/sapcontrol -prot NI_HTTP -nr 02 -function Stop 400

12.07.2023 11:33:14
Stop
OK
Waiting for stopped instance using: /usr/sap/RH2/SYS/exe/hdb/sapcontrol -prot NI_HTTP -nr 02 -function WaitforStopped 600 2


12.07.2023 11:33:30
WaitforStopped
OK
hdbdaemon is stopped.

sidadm+@node3% HDB stop
hdbdaemon will wait maximal 300 seconds for NewDB services finishing.
Stopping instance using: /usr/sap/RH2/SYS/exe/hdb/sapcontrol -prot NI_HTTP -nr 02 -function Stop 400

12.07.2023 11:33:14
Stop
OK
Waiting for stopped instance using: /usr/sap/RH2/SYS/exe/hdb/sapcontrol -prot NI_HTTP -nr 02 -function WaitforStopped 600 2


12.07.2023 11:33:30
WaitforStopped
OK
hdbdaemon is stopped.

Copy to Clipboard

Toggle word wrap

检查节点 3 上的主数据库：

sidadm+@node3% hdbnsutil -sr_stateConfiguration| grep -i "primary masters"                                                node3: Wed Jul 12 11:20:51 2023

primary masters: node2

sidadm+@node3% hdbnsutil -sr_stateConfiguration| grep -i "primary masters"                                                node3: Wed Jul 12 11:20:51 2023

primary masters: node2

Copy to Clipboard

Toggle word wrap

检查集群中的当前主要信息：

root@node1# pcs+ resource | grep Promoted
    * Promoted: [ node1 ]

root@node1# pcs+ resource | grep Promoted
    * Promoted: [ node1 ]

Copy to Clipboard

Toggle word wrap

检查 sr_state 以查看 SAP HANA 系统复制关系：

sidadm@node1%+ hdbnsutil -sr_state

System Replication State
~~~~~~~~~~~~~~~~~~~~~~~~

online: true

mode: primary
operation mode: primary
site id: 2
site name: DC1

is source system: true
is secondary/consumer system: false
has secondaries/consumers attached: true
is a takeover active: false
is primary suspended: false

Host Mappings:
~~~~~~~~~~~~~~

node1 -> [DC3] node3
node1 -> [DC1] node1
node1 -> [DC2] node2


Site Mappings:
~~~~~~~~~~~~~~
DC1 (primary/primary)
    |---DC3 (syncmem/logreplay)
    |---DC2 (syncmem/logreplay)

Tier of DC1: 1
Tier of DC3: 2
Tier of DC2: 2

Replication mode of DC1: primary
Replication mode of DC3: syncmem
Replication mode of DC2: syncmem

Operation mode of DC1: primary
Operation mode of DC3: logreplay
Operation mode of DC2: logreplay

Mapping: DC1 -> DC3
Mapping: DC1 -> DC2
done.

sidadm@node1%+ hdbnsutil -sr_state

System Replication State
~~~~~~~~~~~~~~~~~~~~~~~~

online: true

mode: primary
operation mode: primary
site id: 2
site name: DC1

is source system: true
is secondary/consumer system: false
has secondaries/consumers attached: true
is a takeover active: false
is primary suspended: false

Host Mappings:
~~~~~~~~~~~~~~

node1 -> [DC3] node3
node1 -> [DC1] node1
node1 -> [DC2] node2


Site Mappings:
~~~~~~~~~~~~~~
DC1 (primary/primary)
    |---DC3 (syncmem/logreplay)
    |---DC2 (syncmem/logreplay)

Tier of DC1: 1
Tier of DC3: 2
Tier of DC2: 2

Replication mode of DC1: primary
Replication mode of DC3: syncmem
Replication mode of DC2: syncmem

Operation mode of DC1: primary
Operation mode of DC3: logreplay
Operation mode of DC2: logreplay

Mapping: DC1 -> DC3
Mapping: DC1 -> DC2
done.

Copy to Clipboard

Toggle word wrap

我们需要此信息才能将复制状态与故障转移完成后的状态进行比较。

在集群中启动故障转移，移动 SAPHana-clone-ressource 示例：

root@node1+# pcs resource move SAPHana_RH2_02-clone
Location constraint to move resource 'SAPHana_RH2_02-clone' has been created
Waiting for the cluster to apply configuration changes...
Location constraint created to move resource 'SAPHana_RH2_02-clone' has been removed
Waiting for the cluster to apply configuration changes...
resource 'SAPHana_RH2_02-clone' is promoted on node 'node2'; unpromoted on node 'node1'

root@node1+# pcs resource move SAPHana_RH2_02-clone
Location constraint to move resource 'SAPHana_RH2_02-clone' has been created
Waiting for the cluster to apply configuration changes...
Location constraint created to move resource 'SAPHana_RH2_02-clone' has been removed
Waiting for the cluster to apply configuration changes...
resource 'SAPHana_RH2_02-clone' is promoted on node 'node2'; unpromoted on node 'node1'

Copy to Clipboard

Toggle word wrap

清除资源：

root@node1#  pcs+ resource clear SAPHana_RH2_02-clone
root@node1# pcs resource cleanup SAPHana_RH2_02-clone
Cleaned up SAPHana_RH2_02:0 on node1
Cleaned up SAPHana_RH2_02:1 on node2

root@node1#  pcs+ resource clear SAPHana_RH2_02-clone
root@node1# pcs resource cleanup SAPHana_RH2_02-clone
Cleaned up SAPHana_RH2_02:0 on node1
Cleaned up SAPHana_RH2_02:1 on node2

Copy to Clipboard

Toggle word wrap

检查集群的 sync_state ：

root@node1# pcs status --full | grep sync_state
    * hana_rh2_sync_state             	: SOK
    * hana_rh2_sync_state             	: PRIM

root@node1# pcs status --full | grep sync_state
    * hana_rh2_sync_state             	: SOK
    * hana_rh2_sync_state             	: PRIM

Copy to Clipboard

Toggle word wrap

根据数据库的大小，您需要等待 5 分钟或更多，除非其如下所示。另请参阅检查复制状态。

检查 node3 上的主节点，这应该保持不变：

sidadm+@node3% hdbnsutil -sr_stateConfiguration| grep -i "primary masters"
primary masters: node1

sidadm+@node3% hdbnsutil -sr_stateConfiguration| grep -i "primary masters"
primary masters: node1

Copy to Clipboard

Toggle word wrap

然后，检查新主上的复制状态：

sidadm@node1%+ hdbnsutil -sr_state

System Replication State
~~~~~~~~~~~~~~~~~~~~~~~~

online: true

mode: syncmem
operation mode: logreplay
site id: 2
site name: DC1

is source system: false
is secondary/consumer system: true
has secondaries/consumers attached: false
is a takeover active: false
is primary suspended: false
is timetravel enabled: false
replay mode: auto
active primary site: 1

primary masters: node2

Host Mappings:
~~~~~~~~~~~~~~

node1 -> [DC3] node3
node1 -> [DC1] node1
node1 -> [DC2] node2


Site Mappings:
~~~~~~~~~~~~~~
DC2 (primary/primary)
    |---DC1 (syncmem/logreplay)

Tier of DC2: 1
Tier of DC1: 2

Replication mode of DC2: primary
Replication mode of DC1: syncmem

Operation mode of DC2: primary
Operation mode of DC1: logreplay

Mapping: DC2 -> DC1
done.

sidadm@node1%+ hdbnsutil -sr_state

System Replication State
~~~~~~~~~~~~~~~~~~~~~~~~

online: true

mode: syncmem
operation mode: logreplay
site id: 2
site name: DC1

is source system: false
is secondary/consumer system: true
has secondaries/consumers attached: false
is a takeover active: false
is primary suspended: false
is timetravel enabled: false
replay mode: auto
active primary site: 1

primary masters: node2

Host Mappings:
~~~~~~~~~~~~~~

node1 -> [DC3] node3
node1 -> [DC1] node1
node1 -> [DC2] node2


Site Mappings:
~~~~~~~~~~~~~~
DC2 (primary/primary)
    |---DC1 (syncmem/logreplay)

Tier of DC2: 1
Tier of DC1: 2

Replication mode of DC2: primary
Replication mode of DC1: syncmem

Operation mode of DC2: primary
Operation mode of DC1: logreplay

Mapping: DC2 -> DC1
done.

Copy to Clipboard

Toggle word wrap

"Site Mapping"部分中缺少 DC3，因此除非您将 DC3 重新注册到新主，否则您将无法启动它。

DC3 可以使用以下方法重新注册：

sidadm@node3% hdbnsutil -sr_register --remoteHost=node2
--remoteInstance=${TINSTANCE} --replicationMode=syncmem --name=DC3
--remoteName=DC1 --operationMode=logreplay --online

sidadm@node3% hdbnsutil -sr_register --remoteHost=node2
--remoteInstance=${TINSTANCE} --replicationMode=syncmem --name=DC3
--remoteName=DC1 --operationMode=logreplay --online

Copy to Clipboard

Toggle word wrap

将立即创建缺少的条目，系统复制将在 SAP HANA 数据库启动后立即启动。

您可以通过执行以下内容来检查它：

sidadm@node1% hdbnsutil -sr_state
sidadm@node1% python systemReplicationStatus.py ; echo $?

sidadm@node1% hdbnsutil -sr_state
sidadm@node1% python systemReplicationStatus.py ; echo $?

Copy to Clipboard

Toggle word wrap

您可以在 Check SAP HANA System Replication status 中找到更多信息。

5.4. 测试 2：使用被动第三个站点的主节点故障切换

学习

尝试、购买和销售

社区

关于红帽文档

让开源更具包容性

關於紅帽

Theme

Red Hat legal and privacy links

Red Hat legal and privacy links