I have zookeeper servers collocated with other services on same physical disk, so i have been experiencing disk i/o issues badly for the past few weeks. Increased timeout and rpc interval for both namenode and zookeeper... the failover is happening less frequently as compared to before. I am only using NN HA & RM HA with zookeeper, other than that i do not have any active jobs for kafka and hbase. I would like to know if this workaround(timeout) will keep this cluster good and going with failovers happening until we have a dedicated disk for zookeeper ? should i foresee any other issues in this case?