Yes, attempting to read a huge supercolumn would do it. I'd delete
the supercolumn and compact. You'll need your large heap to do the compaction.
2011/6/24 박상길 :
> When I set the max heap size below 12GB in this node, the cassandra was down
> by OOM, or nearly down - process is not dead but sen
When I set the max heap size below 12GB in this node, the cassandra was down by
OOM, or nearly down - process is not dead but sending with thrift or gossip
resulted in timeout.
When I increased the max heap to 25GB (the node has 32GB of physical memory),
it did not go to OOM, but still got timeo
