Hi, in another thread, I has mentioned that we had issue with Cassandra getting killed by kernel due to OOM. Downgrading to jdk6_37 seems to have fixed it.
However, even now, after every couple of hours, the nodes are showing a spike in memory usage. For ex: on a 8GB ram machine, once the usage reached to 7.5 GB. and then slowly it comes down to normal. Cassandra version in use is 1.1.9.10. Any idea why this could be happening? There is no load on the cluster. Thanks.
