Integration of SUSE Linux Enterprise High Availability Extension and Red Hat Enterprise Linux 7.x and 8.x High Availability Clusters with SAP Landscape Management 3.0
Starting with SAP Landscape Management (LaMa) Support Package (SP) 21 the cluster integration for the SUSE Linux Enterprise high availability extension was implemented. Now with LaMa SP25 it also integrates with Red Hat Enterprise Linux High Availability Clusters, which means that the pacemaker support is completed for all SAP HANA database environments. This blog describes how to configure the integration, what is supported and what are the restrictions with the new features.
Maybe you already found my old post related to this topic. If you use the cluster integration described in this blog post you can consider the previous posts as obsolete. They are not necessary anymore.
Part 1: https://blogs.sap.com/2016/11/19/integration-of-suse-linux-enterprise-high-availability-extension-12-with-sap-landscape-virtualization-management-2.1
Part 2: https://blogs.sap.com/2019/04/24/integration-of-suse-linux-enterprise-high-availability-extension-12-with-sap-landscape-management-3.0
Why should you use the Cluster Manager Extension of SAP LaMa with your pacemaker cluster?
The new feature combines the advantages of both solutions together in one tool. The high availability extension provides you the opportunity to ensure the continuity of your SAP applications and daily business including monitoring. But SAP LaMa enhances this with the capability to also ensure continuity of your operations, administration and maintenance task together with your whole SAP landscape combined in one single point of entry. You will be able to use the automation for almost all administrative tasks available in SAP LaMa
What are the prerequisites to use the Cluster Manager Extension of SAP LaMa?
- You have setup a Red Hat Enterprise Linux or SUSE Linux Enterprise high availability extension cluster managed by pacemaker.
- You have installed SAP Landscape Management, enterprise edition SP25 Patch 0 or later
- You have installed SAP Adaptive Extensions 1.0 patch level 73 or later on all hosts within your cluster. For more information, see SAP Adaptive Extensions, Installation Guide.
- You have installed at least Python version 2.7 or 3.6 and /usr/bin/python should be a symlink to the Python executable on all hosts within your cluster.
- You use SAP HANA databases, which are not running in any scale-out scenario.
Info: According to the official SAP LaMa Road Map, the support for scale-out systems is planned for the upcoming year. See the Road Map Explorer:
How to enable the cluster integration of SAP LaMa?
If you fulfill the prerequisites you can go to Setup –> Settings –> Engine and enable the checkbox “Enable Cluster Integration” –> enable.
Once you enabled the cluster integration SAP LaMa will be aware of SAP systems, which run in a pacemaker cluster. You will be able to run Cluster Manager Operations operations, view cluster details in the operations view when opening the general instance or host details and use cluster validations.
For more information, see Managing Cluster Integration
How does the cluster integration of SAP LaMa work and what is supported?
In principle, you can distinguish between different categories how the cluster is integrated.
|Integration with SAP cluster connector||For ASCS and ERS instances, you have installed and configured the SAP cluster connector (SAP HAlib).
Red Hat: https://access.redhat.com/solutions/3606101
|Start SAP ASCS Instance|
|Operation is automatically adapted||None||Forced Start|
|Operation is no longer executable||The operation is no longer needed, because e.g. the cluster is executing the operation.||Activate Virtual Tenant Addresses|
|You must enable maintenance mode||Before the execution you have to enable the maintenance mode to avoid confusion of the cluster.||Deactivate (OS Shutdown)|
For a full list of all operations, see Managing Cluster Integration
Which cluster validations are available?
SAP LaMa offers some standard validations for the monitoring of the health of your pacemaker cluster. Those validations run on instance and host level and provide you useful information about the state of your cluster.
|ClusterInfo||Shows the identifier for the resource or node in the pacemaker cluster.|
|ClusterMM||Checks if the pacemaker cluster is in maintenance mode.|
|ClusterStatus||Shows the current status for the cluster.|
For a full list of all validations, see Validations for Instances
Which specific operations for the pacemaker cluster are supported?
SAP LaMa integrates with all atomic operations of the pacemaker cluster, so that you can easily manage your high availability cluster from the SAP LaMa UI and don’t have to login to your cluster server.
|Operations for instances||Description|
|Clearing Cluster Resources||Clear the constraints of a resource.|
|Clearing Colocated Cluster Resources||Clear the constraints of a colocated cluster resource.|
|Disabling Cluster Maintenance Mode on Instance||After your maintenance work is done, you disable the maintenance mode for the cluster.|
|Disabling Cluster Resource Maintenance Mode||After your maintenance work is done, you disable the maintenance mode for the cluster resource.|
|Enabling Cluster Maintenance Mode on Instance||To perform maintenance tasks, you enable the maintenance mode for the cluster.|
|Enabling Cluster Resource Maintenance Mode||To perform maintenance tasks, you enable the maintenance mode for the cluster resource.|
|Getting the Cluster Status||View the status of the cluster. You get a status overview of the cluster, cluster nodes, and resources running on the specific nodes.|
|Moving Cluster Resources||Move cluster resources.|
|Moving Colocated Cluster Resources||Move colocated cluster resources.|
|Refreshing Cluster Resources||Delete cluster resource’s history (including failures) so its current state is rechecked.|
|Refreshing Colocated Cluster Resources||Delete colocated cluster resource’s history (including failures) so its current state is rechecked.|
|Starting Cluster Resources||Start cluster resources.|
|Starting Colocated Cluster Resources||Start colocated cluster resources.|
|Stopping Cluster Resources||Stop cluster resources.|
|Stopping Colocated Cluster Resources||Stop colocated cluster resources.|
For a full list of all instance operations, see Cluster Manager Operations on Instances
|Operations for hosts||Description|
|Disabling Cluster Maintenance Mode on Host||After your maintenance work is done, you disable the maintenance mode for the cluster.|
|Disabling Cluster Node Maintenance Mode||After your maintenance work is done, you disable the maintenance mode for the cluster node.|
|Enabling Cluster Maintenance Mode on Host||To perform maintenance tasks, you enable the maintenance mode for the cluster.|
|Enabling Cluster Node Maintenance Mode||To perform maintenance tasks, you enable the maintenance mode for the cluster node.|
|Getting the Cluster Node Status||View the status of the cluster node.|
|Setting Cluster Nodes to Online||Set cluster nodes to online.|
|Setting Cluster Nodes to Standby||Set cluster nodes to standby.|
For a full list of all host operations, see Cluster Manager Operations on Hosts
Which SAP LaMa processes are executable and how are they integrated?
All of the important processes in SAP LaMa designed for the SAP HANA databases are integrated with the cluster environment. All these processes are started in LaMa and even if the cluster is executing the steps it is LaMa, which is triggering all steps of the process. Below you can find a few simplified flow chart examples.
Near Zero Downtime Takeover
This process is thought for a planned takeover, which allows you to switch between primary and secondary database with almost no downtime for the end users logged into the application.
For more information, see Near Zero Downtime Takeover
In a cluster environment the pacemaker cluster will take care of unplanned takeover, due to failures, if you configured your cluster for automatic takeover. But of course you can also use this process to trigger a takeover manually if a failure occurs.
For more information, see Takeover
Near Zero Downtime SAP HANA Update on Primary Tier
LaMa is providing two processes to automatically update your SAP HANA Databases decreasing the necessary downtime. This process is available for primary and secondary SAP HANA database servers.
For more information, see Near Zero Downtime SAP HANA Update on Primary Tier
Near Zero Downtime Maintenance on Primary Tier
In case e.g. hardware maintenance is necessary for one of your database hosts within the cluster this process provides you the opportunity to keep your downtime as minimal as possible. This process is available for both SAP HANA database servers.
For more information, see Near Zero Downtime Maintenance on Primary Tier
What are the restrictions using the cluster integration of SAP LaMa?
The most important restriction is that you must not use systems running within a cluster for any provisioning scenario. Provisioning (clone, copy and refresh) and relocation scenarios are not supported. This means that the whole system provisioning is disabled by default if the cluster integration is enabled and a system is running in an active pacemaker cluster. It doesn’t matter if the system, which is running in a pacemaker cluster, is the source or the target system of your provisioning scenario.
LaMa will show you the following error message once you try to enable a provisioning scenario for a clustered system.
But even if you configured the provisioning scenario before enabling the cluster integration SAP LaMa will show you the same error message in the Provisioning UI.
What to do if you want to use a clustered system in a provisioning process?
Good news first, according to the official SAP LaMa Product Road Map the support for SAP system refresh capabilities (replication based and restore based) for SAP systems managed by the pacemaker cluster manager is planned for the upcoming year. See the Road Map Explorer:
As of SP25 and the standard LaMa configuration it is not possible to allow the provisioning scenarios for systems running in a pacemaker cluster. But you do have two options to workaround this restriction for now. I would recommend to use the first option if you want to use system provisioning for clustered systems.
Recommended Option 1: Disable SAP LaMa cluster integration during provisioning
This option is in my opinion recommended for easier scenarios. Most customers are running their production environment in a cluster and all other systems are not clustered. So if you want to refresh for example your quality system from a production system using replication or restore a backup this option would be the best choice for you. The effort to disable the cluster integration is very low and you can still make use of the standard provisioning scenarios. After the provisioning you can enable the cluster integration again.
Option 2: Build your own Cluster Manager Extension
As a consultant, I would not recommend this option, because it causes a high effort to implement the whole functionality. It would only make sense to choose this way in case the standard provisioning scenario would not be feasible for you at all and you anyhow need a lot of customization in your provisioning process.
If you need some guidance how to realize your own integration you can read my last blog about the pacemaker cluster.
As this would be your own custom implementation you will be responsible for the implementation and it is not possible to get support through the official SAP support channels.
Blog Post: https://blogs.sap.com/2019/04/24/integration-of-suse-linux-enterprise-high-availability-extension-12-with-sap-landscape-management-3.0
How to troubleshoot failed Cluster Manager operations?
First of all in case an operation fails it is very important to determine, which component caused the error. The message code of the SAP LaMa cluster connector is “OS Command” as you can see on the screenshot below.
Besides the standard SAP LaMa operations logs in the monitoring menu, you can also find additional logs directly on the managed host in the SAP Host Agent working directory.
To sum up, the new SAP LaMa cluster integration provides you very useful capabilities to optimize your system orchestration also for high availability landscapes. I hope this blog post could answer some of your questions and support you in the configuration of the cluster integration. Since SP25 is available for download since November 28th, go check it out and see for yourself. In case you still have open questions, please use the comment function to provide us feedback.