On Fri, Nov 27, 2015 at 2:52 AM, Vasiliy I Ozerov <voze...@news360.ru> wrote:
> We have some strange troubles with cassandra startup. Cluster consists of > 4 nodes. 32 Gb RAM per node, each node has about 30Gb of data, 8 CPU. > > So, just after start it has 2753202 pending readstage tasks. And it takes > about 11 hours to complete them all. > I'd presume that it's doing some wacky cache warming, but have no idea what? =Rob