Hello Cloudera Forums,
IT recently cloned our VMs for one of our old Hadoop stacks onto some fancy new hardware and bumped up the specs on the VMs. We had been getting "out of memory" errors on Impala Daemon so IT provided us with a significant boost to the available RAM on each server (I know that we SHOULD have just added on another node but unfortunately that's just how my org operates). Anyways we modified the configuration for the services to use more of the RAM on the VMs but now we are getting the below error:
Memory on host XXXX is overcommitted. The total memory allocation is 223.8 GiB bytes but there are only 252.3 GiB bytes of RAM (50.5 GiB bytes of which are reserved for the system). Visit the Resources tab on the Host page for allocation details. Reconfigure the roles on the host to lower the overall memory allocation. Note: Java maximum heap sizes are multiplied by 1.3 to approximate JVM overhead.
I'm a bit confused as to why we are getting this error, when I increase the RAM used by Impala I still get this error (the system reserving 50Gb) but when I decrease it, the number doesn't change either. I'm a bit confused as to what the issue is. Is there a shared role across the nodes that needs a mirrored config or something?
