[Condor-users] DAGman duplicating jobs on schedd restart

[Christopher Martin <cmartin@xxxxxxxxxxxxxxxxxxxxxx>]

Whenever the schedd restarts we're getting duplicate jobs showing up in the queue. For example if we have a DAG like the following:

JOB A

JOB B

JOB C

JOB D

PARENT A CHILD B

PARENT B CHILD C D

Before the schedd restart, jobs A and B have completed and jobs C and D are queued. After the schedd restarts we then have C and D still queued but B has been added back into the queue as well. Is this a peculiarity of the DAG rescue or perhaps it could be a conflict with the dagman logs?

Any help or leads would be appreciated.

Cheers,

Chris.