IZ20: new testsuite test for Gridengine issue 380

[joga]

[Imported from gridengine issuezilla ​http://gridengine.sunsource.net/issues/show_bug.cgi?id=20]

        Issue #:      20              Platform:     All           Reporter: joga (joga)
       Component:     testsuite          OS:        All
     Subcomponent:    tests           Version:      current          CC:    None defined
        Status:       NEW             Priority:     P3
      Resolution:                    Issue type:    TASK
                                  Target milestone: milestone 1
      Assigned to:    issues@testsuite
      QA Contact:     joga
          URL:
       * Summary:     new testsuite test for Gridengine issue 380
   Status whiteboard:
      Attachments:

     Issue 20 blocks:
   Votes for issue 20:     Vote for this issue


   Opened: Mon Nov 13 01:58:00 -0700 2006 
------------------------


This was IZ 474 in project gridengine.

When errors occured in the startup phase of qrsh
jobs, no error message was reported and the job
was reported as having exited regularly with exit
code 1.

This has been fixed (see Issue 380).

A testsuite test for this issue should do the
following tests:
1. Test regular (successful) job execution with qrsh

2. Set active_jobs directory to unreadable to
others (chmod o-rx active_jobs), submit a qrsh job
under a user id other than the admin user.
An error has to be output on stderr, but no error
should appear in the qmaster messages file, the
queue shall not be set into error state.

3. Execute an unknown command (with and without
qrsh option -noshell) - this is a user error that
shall be output on stderr, but queue will not be
set into error state.

4. Submit qrsh with option -cwd in a working
directory that is not available on the execution
host. User error, see 3.

5. Install the following prolog:
#!/bin/sh
chmod o-r $SGE_JOB_SPOOL_DIR/environment

and submit a qrsh job with a userid other than
admin user. Expect the following behaviour:
- error message in qrsh stderr
- queue in error state
- entry in qmaster messages file
- if abort mail was requested (-M <mailto> -m a),
it should contain the error message
- if adminuser_mail was configured, it should
contain the shepherds trace file including the
qrsh error message

For all tests, check if the accounting entries are
correct.