I don't know if this is true in the latest versions but when we first started running storm we found that it likes its cwd to be the installation location (in your example /etc/storm). When we run under supervisord we actually run a script that does cd to storm installation directory, then runs the storm services. But usually this would be noticeable with an error in the supervisor or nimbus log. You should check those to see if there are any problems.
On Thu, Dec 18, 2014 at 1:30 PM, Sa Li <[email protected]> wrote: > > Hi, All > > > I have 2 questions about supervisor nodes in storm cluster, I have run the > https://github.com/intel-hadoop/storm-benchmark for the storm performance > test. I found grep and trident topologies will kill supervisor (work) nodes > when the test automatically kill the topoplogies after 5 mins. > > 1. My question is, kill a topology might cause the supervisor node be > killed as well? > > 2. Reading storm's official documentation: "As described in Setting up a > Storm cluster, the Nimbus and Supervisor daemons must be run under > supervision using a tool like daemontools or monit. So if the Nimbus or > Supervisor daemons die, they restart like nothing happened." I have > installed the supervisord in cluster (nimbus and worker nodes) > > [program:storm-nimbus] > command=/etc/storm/bin/storm nimbus & > user=stuser > autostart=true > autorestart=true > startsecs=10 > startretries=999 > log_stdout=true > log_stderr=true > logfile=/var/log/storm/nimbus.out > logfile_maxbytes=20MB > logfile_backups=10 > > [program:storm-supervisor] > command=/etc/storm/bin/storm supervisor & > user=stuser > autostart=true > autorestart=true > startsecs=10 > startretries=999 > log_stdout=true > log_stderr=true > logfile=/var/log/storm/supervisor.out > logfile_maxbytes=20MB > logfile_backups=10 > > [program:storm-ui] > command=/etc/storm/bin/storm ui & > user=stuser > autostart=true > autorestart=true > startsecs=10 > startretries=999 > log_stdout=true > log_stderr=true > logfile=/var/log/storm/ui.out > logfile_maxbytes=20MB > logfile_backups=10 > > However, it just exit all the time, see the log > > root@DO-mq-dev:/var/log/supervisor# tail supervisord.log > 2014-12-17 16:41:24,461 INFO exited: storm-nimbus (exit status 1; not > expected) > 2014-12-17 16:41:24,556 INFO exited: storm-supervisor (exit status 1; not > expected) > 2014-12-17 16:41:29,563 INFO spawned: 'storm-supervisor' with pid 349 > 2014-12-17 16:41:29,565 INFO spawned: 'storm-ui' with pid 350 > 2014-12-17 16:41:29,566 INFO spawned: 'storm-nimbus' with pid 351 > 2014-12-17 16:41:29,599 INFO exited: storm-ui (exit status 1; not expected) > 2014-12-17 16:41:31,566 WARN received SIGTERM indicating exit request > 2014-12-17 16:41:31,566 INFO waiting for storm-supervisor, storm-ui, > storm-nimbus to die > 2014-12-17 16:41:31,567 INFO stopped: storm-nimbus (terminated by SIGTERM) > 2014-12-17 16:41:32,569 INFO stopped: storm-supervisor (terminated by > SIGTERM) > > > Any idea what is the problem? > > thanks > > Alec >
