AWS DevOps Blog
Accelerate your Automation for SAP on AWS Operations using Amazon Q Developer
Based on discussions with several SAP on AWS customers, we have discovered that the number of SAP administration or operational task requirements often exceed the capacity of the available team. Due to lack of time, resources, and heavy focus on operations, strategic initiatives like digital transformations often remain unaddressed. Although 1P and 3P automation solutions are available, many organizations do not adopt them, due to cost, internal processes, complexities associated with managing multiple-vendor tooling, etc. While some SAP BASIS teams have successfully automated some tasks, the level of effort and skill set to develop custom scripts is not widely available, in some cases due to a skills gap or insufficient knowledge in scripting. In this blog post we will use Amazon Q Developer, a generative AI coding assistant, and use natural language to create SAP operational automation in a more productive fashion.
Walkthrough
Amazon Q Developer acts as a bridge between logical comprehension and practical coding implementation. It enables SAP BASIS administrators to translate their operational understanding into code by interpreting their logic, articulated in natural language. This approach allows us to accelerate the development process of automation scripts, democratizing script development to a broader base of infrastructure and application administrators. In this case, Amazon Q provides coding suggestions by converting natural English language explanations of logic into operational code, such as an automation script for the operational activity (e.g., Start and Stop of SAP).
The solution is orchestrated in two stages:
- Administrators use Q Developer using natural language to formulate a shell script to perform start and stop operations on a single Amazon EC2 instance.
- Q Developer validates inputs, assessment of system installation, and execution of start/stop commands.
Prerequisites
For the walkthrough, we are using VS Codium for our integrated development environment (IDE) with the latest Amazon Q Developer extension installed. However, you may use any of the supported IDEs.
Prior to starting, it may be important to model the entire workflow. For example, the script may need a number of conditions, checks, and logical considerations. However, for the purposes of our scenario, we focus on three specific conditions, checks, and logical processes. For your specific use case, we recommend incorporating additional logical steps, if needed.
The script we will write has 3 arguments in order to Start/Stop the SAP System.
- The SAP System ID (SID)
- The SAP Instance Number
- The command ‘start’ or ‘stop’ – will start or stop the SAP system.
To run the script the command should look like the example below:
scriptname.sh <SID> <InstanceNumber> <start/stop>
There are also four conditions, checks, and logic blocks in the script,
- First, check if the command has 3 arguments. If any are missing, the system will not be able to perform the intended action.
- Second, check if the SAP system (SID) we are trying to manage is available in the current EC2 instance.
- Third, the SAP Instance Number is checked in the current EC2 instance.
- Lastly, the script needs to tell the system which command to run, based on the third argument (e.g., start or stop).
Important: Comments in Shell scripts start with a ‘#’ sign, and the arguments are indicated by a ‘$<n>’ format; n being the sequence number of the argument. So, in our case:
<SID> : $1
<InstanceNumber> : $2
<start/stop> : $3
Now that we have established the structure of how to call the script and what arguments we are going to pass, lets write the comments in English to get code recommendations from Amazon Q.
Getting Started
1. In VS Codium, create a ‘New File’ for our script. Assign a file name and make sure the file extension ends with a ‘.sh’ (e.g., startstopsap.sh).
Below is an example of the comments we used for our logic. Copy paste this into the file.
Info: The first line #!/bin/bash tells the system to execute the script using the Bash shell. The rest of the lines tell what the script needs to check, the logic it needs to follow and the commands it needs to run.
#!/bin/bash
#This is a script that is going to start and stop an SAP instance based on the given inputs to the script
#The script will receive 3 inputs. <SID> <InstanceNumber> <start/stop>
#If the script did not get 3 inputs, the script will fail with showing the usage guidance.
#Check if the file "/usr/sap/sapservices" exists. If not, fail.
#We will check if the given SID is installed in this server by searching the SID in the file "/usr/sap/sapservices" If it does not exist, fail, otherwise continue.
#Then we will check if the given instance number is installed in this server by searching the Instance Number in the file "/usr/sap/sapservices”. If it does not exist, fail, otherwise continue.
#If all conditions met, check the third output and if it's start, start the sap system using "sapcontrol -nr InstanceNumber -function Start"
#If all conditions met, see the third output and if it's stop, stop the sap system using "sapcontrol -nr InstanceNumber -function Stop"
#Then wait for 2 minutes for the stop command to complete (if stop)
#Remove the ipcs (if stop) by the command “cleanipc InstanceNumber remove”
#If the third input is not start or stop, fail.
#End the script.2. Type #Check Input and press Enter, Q will start making code suggestions. If it does not, you can manually invoke suggestions with ‘Option+C’ on Mac or ‘Alt+C’ on Windows.
Figure 1 – Amazon Q Developer Suggestions
3. To accept suggested code, either press ‘Tab’ or click on ‘Accept’.
The ‘< 1/2 >’ means that there are two suggestions and you may accept the one that is most appropriate for the scenario. Toggle between the suggestions using right and left arrows on your keyboard.
We will accept the code and then press Enter to move to the next line. As soon as you press the Enter key, the next line of code will be suggested.
Important: Amazon Q Developer is non-deterministic, which means that code suggestions produced may be different from what is shown in the blog. If the suggestions look different for you, you can use the arrows on your keyboard to toggle between recommendations, as shown below.
4. Accept the next block of code and eventually close the IF loop. Press Enter.
Figure 2 – Reviewing Suggestions
5. Based on comments in the file, Q should should have enough context to suggest what needs to be done next. The script should check if the /usr/sap/sapservices file exists.
Figure 3 – Checking dependencies
6. Once you accept the code, Q will propose the next lines. Keep accepting the appropriate lines of code until all required sections are completed. Once the script is ready, it should look similar to what is depicted below. Save the script.
Figure 4 – First part of the script
Figure 5 – Second part of the script
Figure 6 – Third part of the script
7. Go to the EC2 instance hosting SAP and use your local text editor (e.g., vi) to create a file with the “.sh” file extension. Let’s say the file is named SAPStopStart.sh
8. Paste the contents of the code from your file in the IDE.
9. Save the file and add execute permissions to the file by running chmod +x SAPStopStart.sh
10. To run the script, use the appropriate arguments as shown below.
SAPStopStart.sh <SID> <InstanceNumber> <start/stop>
Figure 7 – Running the script
Conclusion
Although in this blog post we used a simple example of starting and stopping an SAP system, Amazon Q Developer can be extended to a broader spectrum of SAP operational scenarios. Q Developer’s capabilities can be used to harness a broad range of SAP-related use cases, such as kernel patching, database patching, and beyond. In addition to code suggestions, Q Developer offers a security scanning feature, which can be used for fortifying application security. Amazon Q Developer is available in Pro and Free Tiers and does not require an AWS Account to get started. For the purpose of this blog, we used the Amazon Q Developer Free Tier. To learn more about Amazon Q Developer, click to go to its product page.