Pivotal Knowledge Base

Follow

GPCRONDUMP "ERROR: An Instance of gpcrondump.py is Already Running (pid xxx)"

Environment

 Product  Version
 Pivotal Greenplum  All
 OS  All Supported OS

Symptom

gpcrondump may report an error like below, even though there is no gpcrondump process running on the server:

gpcrondump.py:default-[ERROR]:-An instance of gpcrondump.py is already running (pid 5456)

Cause

While gpcrondump executes, it creates a gpcrondump.pid file in the $MASTER_DATA_DIRECTORY, which tracks the process ID of the dump process and is used to control more than 1 instance of the gpcrondump process at a time. The process ID ex. 5456, in the error listed above, comes from the gpcrondump.pid file.

Usually after the completion of a gpcrondump process, the gpcrondump.pid file is cleaned up, however, sometimes due to either a server crash or other manually forced aborts, it may not be removed. In such cases, if a fresh gpcrondump process is executed one may see such issues.

Also, please be advised, if you perform a search for the process using ps, it may still list the process with the same ID as reported in the error but now that process ID may have been allocated to some other process. So, please don't kill the current running process for that process ID.

Resolution

  • Verify to make sure that there is no existing "gpcrondump" process with this PID (the number mentioned in the error message).
  • Remove the gpcrondump.pid file from $MASTER_DATA_DIRECTORY.
  • Rerun gpcrondump.

Note: $MASTER_DATA_DIRECTORY is a commonly used variable which holds the path to the Greenplum Database master data directory. By default it is /data/master/gpseg-1, but may be different if changed while initializing the database.

Comments

Powered by Zendesk