Andrew Purtell wrote:
Just had another regionserver go down with a
DroppedSnapshotException out of the flusher.
Usually this is because can't get to HDFS. Was this the case here or
another OOME?
Can we get access to the logs Andrew so we can take a look? (Off-list?).
I saw an OOME on jgray cluster around replay of split logs after another
machine crashed (The pressure of deploying a bunch of new regions
because of recent crash and the replay of logs combined seems to have
brought it on).
You are running hbase TRUNK on hadoop 0.18.x?
St.Ack
