The world is heterogeneous. All though Windows Server is running on 75% of data centers, organizations use various Linux distributions to run enterprise applications. If you are in to middleware you know for a fact how critical these applications are and the need to monitor same very closely to avoid service interruption.
Let’s have a quick look on how to setup a monitoring solutions for your Linux servers with SCOM. I’m going describe the process in a high level.
- Set up the environment for monitoring
- Install & import the SCOM Management Pack for Linux/Unix Monitoring in the SCOM management server
- Install the Operations Manager agent in the server to be monitored.
- Define monitors, rules and tasks an necessary
Step 1 | Setup the environment
Below are a list of prerequisites that you need to configure prior setting up the monitoring.
- SSH should be up and running on the destination server. It’s how SCOM talks to the agent.
- Port 22 (for SSH) & Port 1270 (for SCOM Agent) should be opened on both sides.
- OpenSSL should be up and running to for certificate signing. This is vital if you have couple of SCOM management servers and wish to use a SSH key for authentication
- Configure resource pool & Run As accounts for Linux servers. If you’re not sure about this I’ve provided some great articles in the end to refer courtesy of TechNet
Step 2 | Installing the MP for Linux/Unix
Now you already know that a management pack contains the parameters & functions that are require to monitor a specific application, be it an OS or just an application. So in our scenario we will have to install the latest MP for Linux/Unix monitoring installed and imported to SCOM in order to setup basic health monitoring for Linux. You can download these through here.
Step 3 | Agent installation and Discovery
Now for those who wish to automatically discover the Linux resources can run an Discovery Wizard in SCOM. But remember the linux user assigned for the Run As account should have enough privileges (best if that user is in the sudoers list) for the agent installation. If you want to do a manual agent installation all you have to to is to install the System Center SCX agent which comes as an rpm file in the SCOM installation directory in the Linux server using rpm command. Once you run the discovery against the server it will automatically identify that a proper agent is already in place.
Step 4 | Define your own monitoring criteria
Now that you have installed the agent after a short while you may notice the basic system health data is being reported to SCOM. You can start creating your own custom MP to store overrides for existing monitors or create new ones. What I recommend is to leave the system for couple of hours to get settled in and the to start defining the monitoring subjects.
Below are some great resources that I’ve come across when I setup monitoring for Linux. Hope you will find same useful .