One of the server few days ago used up almost all available memory.
free -m shows the following:
total used free shared buffers cached Mem: 31941 31112 828 1 169 1563 -/+ buffers/cache: 29379 2562 Swap: 1023 0 1023
After did some research, issuing the following commands prompted high memory usage on the slab and sreclaimable for the dentry kernel process:
slabtop vmstat -m
To free up the memory, i used the following:
sync; echo 3 > /proc/sys/vm/drop_caches
and the result:
total used free shared buffers cached Mem: 31941 971 30969 1 7 103 -/+ buffers/cache: 861 31080 Swap: 1023 0 1023
Will this permanently resolve the memory issue? i think it will not but will keep monitoring, suspected it was due to recent use of nixstats which caused this memory issue due to frequest SSL request (curl or libcurl).