Download as pdf or txt
Download as pdf or txt
You are on page 1of 25

Week 5: Tune and Test the Cluster

Unit 1: Designing the Cluster Tests


Designing the cluster tests
Overview of planning the cluster tests

• Designing test cases involves:


o Gathering the requirements
o Assessing the possible failure scenarios
o Walking through the possible outcomes
• In a complex, high availability environment, several types of
failures may occur
o The types of failures range from software crashes up to a loss of
the whole network or SAN infrastructure
• The cluster must be able to safely handle all possible failure
scenarios
• Testing each scenario, comparing it with the requirements,
and documenting the recovery procedure will enable
efficient handling of failures when they occur

open.sap.com Slide 2
Designing the cluster tests
Planning the tests to perform

• There will be several test to be performed


• You must plan tests for every conceivable scenario
• Below are some example tests to be performed on the SAP
HANA SR primary:
o Stop Primary Database on node01
o Stop Primary Database on node02
o Crash Primary Database on node01
o Crash Primary Database on node02
o Crash node01
o Crash node02
• Below are some example tests to be performed on the SAP
HANA SR secondary:
o Stop the Secondary Database on node02
o Crash the Secondary Database on node02
o Crash node02

open.sap.com Slide 3
Designing the cluster tests
Example test procedure document

• A test procedure document should


contain all the information required to
perform the test and assess the
outcome
• This includes:
o Stating which component is being tested
o Description of the test
o The procedure for performing the test
o The recovery procedure
o The expected results

open.sap.com Slide 4
Designing the cluster tests
Walk through an example test – establish the initial cluster state

• The following test is based on the settings:


o PREFER_SITE_TAKEOVER=“true”
o AUTOMATED_REGISTER=“false”
o SBD_STARTMODE=clean
• This test will walk through the primary SAP
HANA database suffering a complete failure
• Check the current cluster status
o The cluster is healthy
o All resources are running
o The primary SAP HANA database is active on
node01
• You have now established the initial state
of the cluster and the configuration options
• Have the documentation for this test
available

open.sap.com Slide 5
Designing the cluster tests
Walk through an example test – execute the test procedure

• Create the failure by killing the primary


database
o In a terminal session on the node running the
SAP HANA primary, switch to the <SID>adm
user
o Run: HDB kill-9
o Monitor the progress

open.sap.com Slide 6
Designing the cluster tests
Walk through an example test – monitor the failover process

• The primary SAP HANA database running


on node01 has just been killed
o The multi-state resource
msl_SAPHana_HA1_HDB00 is in an unknown
state
o There is no master (primary) for this resource
o The IP address resource rsc_ip_HA1_HDB00
has moved to node02

• The cluster is in the process of re-


configuring and the SAP HANA primary
database is not located on either node01 or
node02

open.sap.com Slide 7
Designing the cluster tests
Walk through an example test – monitor the failover process

• The resource msl_SAPHana_HA1_HDB00 is


starting the master (SAP HANA primary) on
node02
o The IP address resource
rsc_ip_HA1_HDB00 previously moved to
node02 is active

• The resource msl_SAPHana_HA1_HDB00 is


running on node02
o The SAP HANA primary database is now active
on node02

open.sap.com Slide 8
Designing the cluster tests
What next? Review the test

• The failover of the SAP HANA primary database to node02 is


complete
o From a client, test the connection to the SAP HANA database is
working
• Check the documented expected outcome of the test
• Compare the documented outcome with the actual outcome
• Make any required updates:
o Documentation
o Cluster configuration
o Operations Manual

open.sap.com Slide 9
Thank You!

Contact Information:
open@sap.com
Week 5: Tune and Test the Cluster
Unit 2: Activating SAP HANA SR on the Cluster
Activating SAP HANA SR on the cluster
Demonstration – activate SAP HANA SR on the cluster

• The cluster configuration is complete


o The cluster has been in maintenance mode during the
configuration process
• The SAP HANA status will be checked on both nodes to
ensure it is stopped
• The cluster status will be checked
• The cluster will be taken out of maintenance mode
• Hawk will be used to monitor the cluster loading the
resources
• The cluster status will be checked to ensure everything is
running as expected

open.sap.com Slide 2
Activating SAP HANA SR on the cluster
Demonstration summary – activate SAP HANA SR on the cluster

• The cluster configuration is complete, and the cluster has


been in maintenance mode during the configuration process
• SAP HANA was stopped on both nodes
• The cluster status was checked

• The cluster was taken out of maintenance mode

open.sap.com Slide 3
Activating SAP HANA SR on the cluster
Demonstration summary – activate SAP HANA SR on the cluster

• Hawk was used to watch the resources load

• The resources have started but there is a monitor error on the SAP HANA resource on
node02

open.sap.com Slide 4
Activating SAP HANA SR on the cluster
Demonstration summary – activate SAP HANA SR on the cluster

• The monitoring error on the SAP HANA resource on node02 was


cleared
o This is a temporary warning which occurred as the cluster came out of
maintenance mode

• All the resources have started and there are no warnings or errors

open.sap.com Slide 5
Activating SAP HANA SR on the cluster
Demonstration summary – activate SAP HANA SR on the cluster

• SAP HANA with system replication is now active on the


cluster
• The next stage is to perform tests on the cluster
hosting SAP HANA to make sure it performs as the
design specification states

open.sap.com Slide 6
Thank You!

Contact Information:
open@sap.com
Week 5: Tune and Test the Cluster
Unit 3: Testing the Completed Cluster Configuration
Testing the completed cluster configuration
Demonstration – perform an example test on the SAP HANA SR primary

• In this demonstration we will performed a single test on the


cluster hosting SAP HANA SR
o To fully test the cluster other tests must be performed
• The cluster and configuration will be checked before
performing the test
• The SAP HANA primary instance running on node01 will be
stopped using SAP tools and the cluster’s reaction to the
SAP HANA failure will be monitored using Hawk
• The cluster will be monitored until the SAP HANA
primary has been recovered
• The cluster status will be checked using the command line

open.sap.com Slide 2
Testing the completed cluster configuration
Demonstration – perform an example test on the SAP HANA SR primary

• Once the failover has completed, the SAP HANA system


replication configuration on node01 will be viewed
• The SAP HANA system replication status will be checked
• The SAP HANA SR deployment will be returned back to
full HA, the SAP HANA primary will be running on
node02 and the SR secondary on node01

open.sap.com Slide 3
Testing the completed cluster configuration
Demonstration summary – perform an example test on the SAP HANA SR
primary

• In this demonstration we performed a single test on the


cluster hosting SAP HANA
o To fully test the cluster other tests must be performed
• The cluster status before performing the test was
established

• The SAP HANA primary instance running on node01 was


stopped and monitored using Hawk

open.sap.com Slide 4
Testing the completed cluster configuration
Demonstration summary – perform an example test on the SAP HANA SR
primary

• After some time the IP address resource was running on


node02 and the multi-state clone master SAP
HANA resource was also running on node02
o At this point the cluster has recovered the SAP HANA primary and
it is active on node02

• The cluster status was then checked using the command


line
open.sap.com Slide 5
Testing the completed cluster configuration
Demonstration summary – perform an example test on the SAP HANA SR
primary

• The SAP HANA system replication configuration on node01


was checked

• The system replication status was checked from node01

• The next step is to return the SAP HANA SR deployment


back to full HA

open.sap.com Slide 6
Testing the completed cluster configuration
Demonstration summary – perform an example test on the SAP HANA SR
primary

• The SAP HANA SR instance on node01 was configured to be


an SR secondary with node02 as the primary

• To bring back full HA the SAP HANA cluster resource must


be refreshed

open.sap.com Slide 7
Thank You!

Contact Information:
open@sap.com

You might also like