[OODT-765] Stale cas.workflow.pid PID files should be removed or set to blank

• Type: Improvement
• Status: Resolved
• Priority: Minor
• Resolution: Won't Fix
• Affects Version/s: 0.7
• Fix Version/s: 1.1
• Component/s: file manager, resource manager, workflow manager

Environment

Ubuntu 14.04 LTS - 32 bit

• Skill Level: Committer (Medium) - This is for regular contributors/committers
• Assignee: Unassigned
• Reporter: Sabarish Venkatraman
• Created: 26/Oct/14 07:59
• Updated: 16/Jul/17 21:04
• Resolved: 16/Jul/17 21:04

Description

$OODT_HOME/filemgr/run - Contains PID of filemanager
$OODT_HOME/resmgr/run - Contains PID of resourcemanager
$OODT_HOME/workflow/run - Contains PID of workflowmanager

The PIDs are stored in cas.filemgr.pid when these managers are started.

At times, some of these managers might not launch fully due to some internal exceptions, after a PID had been written to these files.

During the subsequent launch, if a PID is already present in the file, it fails to launch the filemgr/workflow manager/resource mgr.

Sample filemgr launch issue:
root@ubuntu:/home/user/Desktop/project2/oodt-deploy/filemgr/bin# ./filemgr start
PID file (/home/user/Desktop/project2/oodt-deploy/filemgr/run/cas.filemgr.pid) found. Is File Manager still running? Start aborted.

The stale process ids stored in the files should be handled.

Fix:
Invalidate the contents of the file by setting it to blank. (or)
Deleting the file upon process termination or during manager launch failure.

Temporary fix:
Delete the cas.workflow.pid file manually and then start the appropriate managers.