hadoop endpoint

["Roberto Maggi @ Debian" <de...@gmail.com>]

Hi you all,

I'm still new to hadoop and I'm still trying to understand a couple of 
things.

I setup a lab with 4 machines, 2 namenodes and 2 datanodes.

One namenode is active active and one in standby.

First question:

I set it up to work with solr but in its configuration I can configure 
only one hadoop endpoint,

becoming the single point of failure.

( from my /opt/solr/bin/solr.in.sh )

...

SOLR_OPTS="$SOLR_OPTS -Dsolr.directoryFactory=HdfsDirectoryFactory"
SOLR_OPTS="$SOLR_OPTS -Dsolr.lock.type=hdfs"
SOLR_OPTS="$SOLR_OPTS -Dsolr.hdfs.home=hdfs://had4:8020/solr"
...


is there an official way to solve this problem or it would be better to 
put them behind a

reverse proxy?

Second question:

since solr cannot use the mapreduce function of hadoop what would be the 
exact
advantage to use hdfs instead of a usual file system?

Thanks in advance


Roberto


---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@hadoop.apache.org
For additional commands, e-mail: user-help@hadoop.apache.org