Question: I’ve received some monitoring alerts about low memory availability on a Linux server hosting a DB2 LUW installation.
I’d like to be able to check myself. What are some indicators I can report on – giving me more detail on the low memory usage?
Also, what are some of the signals that Linux memory health is good?
Answer: Linux offers some commands which will allow you to assess the Linux memory usage. These suggestions are for general usage and it’s important to understand these figures in the context
of the server. A deeper understanding is required to capacity plan and troubleshoot.
When I have a situation of low memory usage some I make some checks to identify if the alerts are valid .
- Swap memory increases (or moves frequently) Read more on Linux swap space and DB2
- Free and buffers/cache = 0 or very close to 0
- dmesg | grep oom-killer --display OutOfMemory-killer ,If there no value returned then oom-killer is not running. The oom-killer sacrifices processes to free memory – it’s a last resort scenario
One of these 3 indicators or all 3 is a strong argument for some immediate action. It may require some troubleshooting to identify root cause .