In this documentation I will explain how to setup and configure a Hana TDI scenario test environment. I will show in detail step and configuration point to achieve this configuration.
For my Hana TDI test case I’ll use my own lab on Vmware Vsphere 5.1.0 and run Hana revision 91 with the following deployment scenario:
- 1 master + 1 worker (load balancing)
- 1 master + 1 worker + 1 standby (HA)
- 1 Hana on Primary site + 1 Hana on Secondary site (DR replication)
Disclaimer: this is a personal documentation for test purpose; I will willingly bypass the mandatory hardware check and HW configuration check
In order execution
- Configure Vsphere for the relevant scenario
- Configure and setup scenario 1 - load balancing
- Configure and setup scenario 2 - High Availability
- Configure and setup scenario 3 - DR replication
Guide used:
SAP Hana Administration Guide SP9
SAP HANA virtualization and TDI_V2
SUSE Linux Enterprise High Availability ExtensionSLEHA
Note used:
1944799 - SAP HANA Guidelines for SLES Operating System Installation
2070786 - SAP systems on VMware vSphere: minimum SAP and OS requirements
1788665 - SAP HANA Support for Virtualized Environments
1943937 - Hardware Configuration Check Tool - Central Note
1969700 - SQL statement collection for SAP HANA
Link used:
Configure Vsphere for the relevant scenario
From the above diagram my esxi server is configure with 4 internal disk, each of them for a specific purpose, I have installed and configured a virtual NAS appliance (freenas) in order to set nfs shared and mount point.
I will explain how to create new SLES vm from a template for Hana so it will be easier to have a reference (image ready) to deploy a new server
SLES installation
- 4 cores
- 32 GB of Ram
- 30 GB of local disk for OS
- 2 NIC card
I obviously did not detail how to create a new vm within vsphere since it's very straight forward, but once the new SLES server is ready
I start some post work in order to have the VMware tools installed and make it Hana SP9 compliant
Now let’s make the system Hana SP9 relevant; in order to do this some additional package needs to be installed on the system regarding the note “1944799 - SAP HANA Guidelines for SLES Operating System Installation”. Those package can be install by using the os command or by using the tool “Yast” ,“Yast2” or “Zypper”In my case this two components were too low, the required version is 17.2
Once deploy I install the specific package for Hana (check the following link for further information on what this package is all about)
My system is all set now; I perform a snapshot of it and create a template for the future deployment
The VM is now ready to have Hana installed on it, but first I need to set the volume used for the installation.
I will use my FreeNas VM appliance to set my volume with the following rule:
/hana/log FS = 0,5 x Ram (shared)
/hana/data FS = 1 x Ram (shared)
/hana/shared FS = 1 x Ram (shared)
/usr/sap FS = 20 GB (local drive)
The drives ready I now add it to my server
Hana is ready now to be installed.
Configure and setup for scenario 1 – Load Balancing
Let start the first scenario by installing Hana single node, before to run the installed I run the HanaHwCheck script to see if my server component are supported for a Hana installation.
Of course it failed!!! but it’s the first thing that needs to be run before doing an install, in order to by passed this check during the install I will set the following variable IDSPISPOPD="1".
Once set do a quick check on the script again
Now the second test performance script will needs to be run to validate TDI approach, to do this first download the “Hardware config check “ tool on the market place and extract it into “/hana/shared” (also available from the complet dvd set)
Form the note “1943937 - Hardware Configuration Check Tool - Central Note” I adapt the .json file setting according my environment (refer to the attach document in the note) and copy them into “/hana/shared/hwcct”Once done I run my performance test
The output tells me what needs to be corrected; I don’t run all of them since it’s for a lab test purpose
The installation of the master node is now done
I order to now install the second node for load balancing I’ll deploy the new server by using VM Template
I fast forward since it’s pretty straight, I did customize the template in order to receive the new config
My new server ready let's add it to the current landscape, but before to run the script to add the new host I first need to set the master parameter “listeninterface” to “global”
Now the new server go to “/hana/shared/HB1/global/hdb/install/bin” and run the script ./hdbaddhost
The new host appears
Paid attention now to the fact that the system in now distributed, let's go to the next scenario.
Configure and setup for scenario 2 – High Availability (basic scenario)
Since I did explain how to install a new vm over a template, I’ll skip this part in the documentation but will focus more on the standby installation and failover test.Quick sum of activities not documented in this section:
- Deployment of vm from template
- Adjustment of parameters
- Filesystem mount for new system
Proceed the same way to add a new host like previously, but choose option "2 - Standby".
Now that the standby node has been added let’s make a failover test to see how Hana react, to force the take over from host03 I’ll perform a HDB stop on the node 2 (vmhana02)
Stop at 4:36 pm, failover finish at 4:38 and we can see that the node #3 become slave and not standby anymore
Since there is no automatic failed back, once the node #2 is back it become the Standby node
For the next scenario, check the second part of the document below.