Checking RAS Status

There are two main methods that can be used to check the health of the RAS system: checking resource usage and checking the log file.

The steps below are all specific to each RAS web server. If the system is running poorly or is down, both servers (raweb-prod-01 and raweb-prod-02) should be checked to determine which server (or both) should be restarted.

Checking Resource Usage

1. SSH to the web server (raprod-web-01, raprod-web-02) using PuTTY or SecureSSH.
   
   

2. Run the 'top' command to display running processes:

   top

3. By default, the processes are sorted by CPU usage, which changes from moment to moment, causing the order of the list to constantly update. To make it easier to read, sort by memory usage:

   1. With top still running press "Shift + F" to bring up the sort screen.

   2. Press the "N" key on your keyboard to select "Memory usage", then hit Enter.

   3. The list will now be sorted by memory usage. The top entry should be a process named "java" -- this is the RAS process.
      
      

4. The important values to note are the CPU usage (%CPU) and Memory usage (%MEM%).

   1. Memory usage should normally be anywhere from 25% to 50%.

   2. CPU usage will go up and down frequently, but should mostly be below 50%.

   Fig. 1: 'Top' screen with CPU and memory info
   
   
   

5. If %MEM is at 80% or more, or if %CPU is consistently above 100%, the RAS process on that server most likely needs to be restarted.
   
   

6. When finished, press the "Q" key on your keyboard to exit the top application.

Checking Log Files

1. SSH to the web server (raprod-web-01, raprod-web-02) using PuTTY or SecureSSH.
   
   

2. Run the following command to stream the contents of the RAS log file to your screen:

   tail -f /usr/local/tomcat7/logs/catalina.out

3. A lot of information is printed to the log file as users use the application. The details aren't important, but errors look very different than normal activity and are usually easy to spot -- see the screenshots below.
   
   If the log is repeatedly and consistently printing blocks of the shorter error lines shown in Fig. 2, the RAS process on that server most likely needs to be restarted.
   
   
4. You can press Ctrl+C to close the log file.

Fig. 1: A normal log file – long, dense lines with the date and time on the far left. (They may contain things like 'WARN' or 'Exception' but that's fine.)

Fig. 2: An error-filled log file: shorter lines indented on the left, and Java filenames (.java) and numbers at the end