Thanks Konstantin. I understand now
" At startup the name node accepts data node reports collecting
information about block locations. In order to leave safe mode it needs
to collect a configurable percentage called threshold of blocks, which
satisfy the minimal replication condition. The minimal replication
condition is that each block must have at least dfs.replication.min
replicas. When the threshold is reached the name node extends safe mode
for a configurable amount of time to let the remaining data nodes to
check in before it will start replicating missing blocks. Then the name
node leaves safe mode."
I realized that lots of blocks are missing their replicas. And that
turns into the situation of SAFE mode.
Does it make sense to allow the command ./hadoop dfs -setrep to work
even the name node is in safe mode? Otherwise the name node is idle forever
Konstantin Shvachko wrote:
You can run "hadoop fsck / " to see how many blocks are missing on
your cluster.
See definition of safe mode here.
http://lucene.apache.org/hadoop/api/org/apache/hadoop/dfs/NameNode.html#setSafeMode(org.apache.hadoop.dfs.FSConstants.SafeModeAction)
--Konstantin
erolagnab wrote:
Hi all,
Just wondering what is the reason causing NameNode is on SafeMode
forever?
I've left my machine running for 2 days and it's still on Safe Mode.
Trung