In any high load system memory management is a important issue. Linux kernel allow to modify the way it allocate virtual memory based multiple factors. These parameters can be controlled and obtain a better performance as the requirement. Linux systems typically let you allocate as much virtual memory as you like, then kill the process if it uses too much of it. Virtual memory allocation is based on mode of overcommitting, and this is decided based on the value set in
0 - heuristic overcommit 1 - always overcommit, never check 2 - always check, never overcommit
Mode 0 - This is the default mode of operation. Calls of mmap(2) with MAP_NORESERVE are not checked and the default check is very weak, leading to the risk of get- ting a process “OOM-killed”.
Mode 1 - This is a disaster when the system is in over loaded mode.
Mode 2 - The total virtual address space on the system is limited to based on virtual memory commit calculation.
Totel Virtual Memory is,
Vm = (Ss +Spm * ( Or / 100 ) )
Ss - Size of the Swap space Spm - Size of the physical memory ( RAM ) Or - Overcommit Ratio
In the above two variables are known and the Overcommit ratio is set in
Os kill the process if it uses too much of memory, this happens can avoid this by switching to mode 2
echo 2 > /proc/sys/vm/overcommit_memory