Troubleshooting starting DataStax Enterprise 

Troubleshooting tips for starting DataStax Enterprise.

To identify and resolve start problem, try these steps before contacting DataStax Expert Support.

Start DataStax Enterprise 

Start DataStax Enterprise as a service or stand-alone process:
Installer type Steps
DSE Services or package installations
$ sudo dse start
DSE No-Services or tarball installations
$ install_location/dse cassandra option 
For options see:5.0, 4.8, 4.7, 4.6

Verify DataStax Enterprise status 

Install type Steps
DSE Services or package installations
$ sudo service dse status
DSE No-Services or tarball installations
$ install_location/nodetool status

Review log files 

If DataStax Enterprise is not running, or starts running and then stops, look for errors at the end of the system log file:
Install type Steps
DSE Services or package installations
$ cat /var/log/cassandra/system.log
DSE No-Services or tarball installations
$ install_location/dse/logs/cassandra/system.log

Discover the Cassandra process ID

It is possible that an active Cassandra process is interfering with starting DataStax Enterprise. For example, on a local development node:

  1. Verify the node status:
    Install type Steps
    DSE Services or package installations
    $ nodetool status
    DSE No-Services or tarball installations
    $ install_location/nodetool status
    Datacenter: Cassandra
    =====================
    Status=Up/Down
    |/ State=Normal/Leaving/Joining/Moving
    --  Address    Load       Owns  Host ID                               Token                Rack
    UN  127.0.0.1  206.23 KB  54%   392b0ca3-3090-440f-8a17-93702234aeec  1093280155279701211  rack1
  2. However, there can be a situation where Cassandra is running, but DataStax Enterprise is not. For example, if your environment had earlier product versions installed or the product was not shut down gracefully, look for the Cassandra process ID and terminate the Cassandra process:
    $ pgrep -f cassandra
    29474
    $ sudo kill -9 29474

Subprocesses not killed when DataStax Enterprise is shut down improperly 

To prevent this problem, avoid using kill 9 and shut down DataStax Enterprise properly.
Install type Steps
DSE Services or package installations
$ nodetool status
DSE No-Services or tarball installations
$ install_location/nodetool status
If DataStax Enterprise is shut down with kill -9, you must restart the node or manually kill any remaining sub-processes:
  • DSE Services or package installations installations:
    For example, if DataStax Enterprise was started using $ sudo services dse start or $ sudo /etc/init.d/dse start and the main process was killed using $ kill -9 `cat /var/run/dse/dse.pid`:
    1. To view the subprocesses left behind (all DSE processes run under user "cassandra"):
      $ pgrep -c -ucassandra >/dev/null && ps -o pid,ppid,user,args `pgrep -ucassandra`
    2. To shut down the subprocesses:
      $ sudo pkill -ucassandra
  • DSE No-Services or tarball installations installations:
    For example, if DataStax Enterprise 5.0 was started using $ sudo dse cassandra -k -t and the main process was killed using $ sudo kill -9 `cat /var/run/dse/dse.pid` or $ sudo pkill -9 -f jmxremote.port=7199:
    1. To view the subprocesses left behind:
      $ pgrep -c -f dse-5.0 >/dev/null && ps -o pid,ppid,user,args `pgrep -f dse-5.0

      All DataStax Enterprise processes run under user cassandra.

    2. To shut down the subprocesses:
      $ sudo pkill -f dse-5.0
Note: The kill command (SIGTERM) shuts down the subprocesses.