Home > Lab 05. Change to Cost Optimized Scenario > Task 04. Test Cluster

Task 04. Test Cluster

Test the changed cluster to verify if it operates normally according to the scenario.

The first scenario is to check how the cluster works when prihana’s DB crashes.

Connect to prihana through Session Manager.
- After log in AWS Management Console, connect to EC2 Instance Console
- Select HANA-HDB-Primary instance, click Action, and click Connect.
- Select Session Manager and click Connect.
- After switching to hdbadm user in prihana node, Crash(HDB kill) HANA DB.
```
sudo su - hdbadm
HDB kill -9
exit
```
- After switching to root user in prihana node, monitor the cluster status.
```
sudo su -
crm_mon -rfn1
```
- Monitor HSR status. Check the status of sechana node is SOK.
```
SAPHanaSR-showAttr
```
- Since prihana instance is in normal state and PREFER_SITE_TAKEOVER option has been changed to false, DB was restarted in prihana without being taken over to sechana.

The second scenario checks how the cluster works when Stop prihana Instance.
- Crash OS(Stop Instance)
  - After log in AWS Management Console, connect to EC2 Instance Console
  - Select HANA-HDB-Primary instance, click Instance state, and click Stop Instance.
- Connect to sechana through Session Manager.
  - After log in AWS Management Console, connect to EC2 Instance Console
  - Select HANA-HDB-Secondary instance, click Action, and click Connect.
  - Select Session Manager and click Connect.
- Monitor the cluster status
```
sudo su -
crm_mon -rfn
```
- Since prihana node was stopped, it was normally taken over by sechana. QAS system was shut down for normal service, and Memory limit setting has been released so that memory can be serviced normally.
```
sudo su - hdbadm
cat /usr/sap/HDB/SYS/global/hdb/custom/config/global.ini
```
Start prihana node. If prihana node is normall, take back to prihana to use QAS system. Restore sechana’s PRD(HDB)’s global.ini settings.
- Start prihana instance.
  - After log in AWS Management Console, connect to EC2 Instance Console.
  - Select HANA-HDB-Primary instance, click Action, Select Start Instance from Instance state.
- Switch to root user in sechana node. After prihana instance starts up normally, check if HDB HANA resource is Slave.
```
sudo su -
crm_mon -rfn
```
- On sechana node, Check HSR status to verify if prihana status is SOK.
```
SAPHanaSR-showAttr
```
- On sechana node, take back to prihana. Change node to standby to move all resources to prihana.
```
crm node standby sechana
```
- Verify prihana’s HDB HANA resource has been switched to Master.
```
crm_mon -rfn
```
- After switching to hdbadm user in sechana node, restore global.ini using VI editor.
```
sudo su - hdbadm

vi /usr/sap/HDB/SYS/global/hdb/custom/config/global.ini

[system_replication]
...
preload_column_tables = false #Add-on

[memorymanager]
global_allocation_limit = 24576
```
- After switching to root user in sechana node, make node online.
```
sudo su -
crm node online sechana
```
- Verify sechana’s HDB HANA resource is changed Slave and QAS HANA resource is Started
```
crm_mon -rfn1
```
- Check the HSR status to verify if sechana status is SOK.
```
SAPHanaSR-showAttr
```