[Condor-users] Can't find WANT_SUSPEND in internal ClassAd

["David A. Kotz" <dkotz@xxxxxxxxxxxxx>]

I'm getting the error in the subject line from some of my machines when 
parallel universe jobs try to start on them.  Any idea what might be 
causing it?  After a few false starts like these, the jobs usually run. 
 All of the machines in the parallel pool are similar and are running 
the same Condor config.

______________________________


This is an automated email from the Condor system
on machine "[machine]".  Do not reply.

"/lusr/condor/sbin/condor_startd" on "[machine]" exited with status 4.
Condor will automatically restart this process in 13 seconds.

*** Last 20 line(s) of file StartLog:
2/21 10:31:41 vm5: Changing state: Owner -> Unclaimed
2/21 10:31:41 vm6: State change: IS_OWNER is false
2/21 10:31:41 vm6: Changing state: Owner -> Unclaimed
2/21 10:31:41 vm7: State change: IS_OWNER is false
2/21 10:31:41 vm7: Changing state: Owner -> Unclaimed
2/21 10:31:41 vm8: State change: IS_OWNER is false
2/21 10:31:41 vm8: Changing state: Owner -> Unclaimed
2/21 10:36:18 DaemonCore: Command received via TCP from host <[IP]:53561>
2/21 10:36:18 DaemonCore: received command 442 (REQUEST_CLAIM), calling 
handler (command_request_claim)
2/21 10:36:18 vm1: Request accepted.
2/21 10:36:18 vm1: Remote owner is 
DedicatedScheduler@[machine].cs.utexas.edu
2/21 10:36:18 vm1: State change: claiming protocol successful
2/21 10:36:18 vm1: Changing state: Unclaimed -> Claimed
2/21 10:36:18 ERROR "Can't find WANT_SUSPEND in internal ClassAd" at 
line 992 in file Resource.C
2/21 10:36:18 vm1: Changing state and activity: Claimed/Idle -> 
Preempting/Killing
2/21 10:36:18 vm1: State change: No preempting claim, returning to owner
2/21 10:36:18 vm1: Changing state and activity: Preempting/Killing -> 
Owner/Idle
2/21 10:36:18 vm1: State change: IS_OWNER is false
2/21 10:36:18 vm1: Changing state: Owner -> Unclaimed
2/21 10:36:18 startd exiting because of fatal exception.
*** End of file StartLog