1) Analysis of services/processes
If you identify that your server has a high CPU, RAM or HDD workload, then it is important to analyze why this is happening. The task manager (Image1-Task manager) can help you check which service/services are responsible for the high workload regardless of whether this is an error or not. The Intrexx processes all have the same name. There are three services by default: a portal, a supervisor and Solr (search server). As these services cannot be told apart based on their names, you need to identify which service is which. To do this, show the PID (Process ID) in the task manager. This PID is a unique identifier of the process (s. Image2-PID). You can then compare the PID of the process with the PIDs of Intrexx and thus, match the corresponding service.
- You can find the PID of a portal in the portal.pid file in the respective portal directory
- You can find the PID of the supervisor in the supervisor.pid file in the installation directory
Example: PID 7228 --> Portal "portal1903" (s. Image3-Match PID)
2) Analyze service/process values
If you have identified a service/process as the culprit, you now need to analyze whether the use of resources is explainable or whether there may be a problem.
It is by all means possible, that a portal service can take up 2GB RAM depending on the scope of the portal, number of users, etc. – and even more in some cases. Based on experience, you can say that portals with up to 50 users should get by with 2GB RAM. However, as the memory required not only depends on the number of users, we cannot make a precise statement at this point. Intrexx processes as well as search indexes or other tasks that are scheduled or started manually can also be memory-intensive or require RAM. You need to check which task or action in the portal is responsible for taking up resources on a case-by-case basis. The Solr service is configured to take up up to ¾ of the physically available memory. Therefore, if there are tasks in the task scheduler (application or file indexes) that are scheduled or are even started manually in parallel, the Solr service can be all means take up as much as 4-6GB RAM. Depending on the scope of the search indexes, memory usage of 5GB for the Solr server is normal; this also applies to test systems.
If the values for your system should be different from the normally seen values (portal service 2GB+ or Solr service (5GB), then you should contact the support team at United Planet or try to identify the problem.
3) Third-party programs and their impact
Portal performance can also greatly depend on the third-party programs, such as the database, virus scanner or other software, which is installed on the server.
You should always bear in mind that the resources on a server have to be divided. Depending on the setup, there are advantages and disadvantages to having the database on the same server as Intrexx. This is usually expressed by the fact that the resources are not all required by Intrexx, but also by the database.
Virus scanners are another important point. This can have a huge impact on the speed of a portal as well as the resources consumed on the server. More information is available in another Knowledge Base article about virus scanners. In principle, if you are having resource problems, you should deactivate or even uninstall these systems in a test environment.