Issues.zip (15.6 KB)
I installed dremio-community-2.0.1rc-201804132202000759_10b1de0_mapr_1.noarch into a single node vm environment with MapR 5.2 installed.
I could not get the Dremio service to start and to load Dremio UI.
Herewith attached are all the logs that I extracted.
Looking at server.out, it reports the following issue
HADOOP_HOME or hadoop.home.dir are not set.
Could you try either set HADOOP_HOME in dremio-env and/or hadoop.home.dir in dremio.conf
Did you try to start Dremio twice by any chance? From the error message in
server.out (Cluster identity not found), it looks like the directory for the internal database was created but left empty, most likely because the first attempt at starting Dremio failed.
I would recommend to remove the directory
/var/lib/dremio/db (it should be empty), restart Dremio and capture the log files which should now show the root cause.
There is no folders or files found in /var/lib/dremio.
The PID file is created in /opt/dremio/run, but each time I check its status, it stated that dremio is not running.
I have set Hadoop.home.dir in dremio.conf already and it still cannot work. You can refer to my dremio.conf in my post attached as Issues.zip.
dremio.conf is actually invalid:
hadoop.home.dir is not a dremio setting but a system property which could be set in
dremio-env, but which is actually okay to leave as-is (the log message is actually a debug message, not an actual error). It seems also you might have changed the logging configuration to be much verbose than usual, and some of these messages might be confusing.
/var/lib/dremio is actually empty, it might be worth checking that the directory is owned by the same user as the one used to run the dremio daemon, which might be
mapr for a default mapr RPM installation, but a different one if MapR impersonation is enabled.