Post running spark jobs os cache memory of cassandra server increases heavily and not getting reduced automatically when i checked i found that buffer/cache is using 10 GB memory i have clean up same using ubuntu drop cache method. ( sync; sudo sh -c "echo 1 > /proc/sys/vm/drop_caches;sync";sudo sh -c "echo 2 > /proc/sys/vm/drop_caches;sync";sudo sh -c "echo 3 > /proc/sys/vm/drop_caches"/ ).
is it okay to clean up os cache on cassandra node ?
Not cleaning this causes outofmemory issue after few days , but cache cleanup working as workaround for this.
Could you please guide about best practices about the same.