We're about to release a big change in DAGMan that will affect output
to the .dagman.out file. Currently when DAGMan submits a job, it forks
a new condor_submit process to do the submission and that output is
logged. In our new approach, DAGMan will directly submit the job to
the condor_schedd without forking a new process. This will result in
different messages.

Currently with the condor_submit forked process, the output looks
something like this:

01/03/22 10:26:28 submitting:
/nobackup/condor/release_dir/bin/condor_submit -a dag_node_name' '='
'SleepA -a +DAGManJobId' '=' '2331 -a DAGManJobId' '=' '2331
-batch-name sleep.dag+2331 -batch-id 2331.0 -a submit_event_notes' '='
'DAG' 'Node:' 'SleepA -a dagman_log' '='
'/scratch/condor/sleep/./sleep.dag.nodes.log -a +DAGManNodesMask' '='
'"0,1,2,4,5,7,9,10,11,12,13,16,17,24,27,35,36" -a JOB=SleepA -a
DAG_STATUS' '=' '0 -a FAILED_COUNT' '=' '0 -a +KeepClaimIdle' '=' '20
-a notification' '=' 'never -a +DAGParentNodeNames' '=' '"" sleep.sub
01/03/22 10:26:28 From submit: Submitting job(s).
01/03/22 10:26:28 From submit: 1 job(s) submitted to cluster 2332.
01/03/22 10:26:28 From submit: WARNING: the line 'JOB = SleepA' was
unused by condor_submit. Is it a typo?
01/03/22 10:26:28 assigned HTCondor ID (2332.0.0)
01/03/22 10:26:28 Just submitted 1 job this cycle...

With the new direct submission method, the output will look like:

01/03/22 11:58:05 Submitting node SleepA from file sleep.sub using
direct job submission
01/03/22 11:58:05 Submit warning: Submit:0:the line 'RETRY = 0' was
unused by DAGMAN. Is it a typo?

This change is not due to go live until our 9.6.0 release at the
earliest, which will release in late February. Also you can easily
revert to the old forked condor_submit approach by setting a
configuration knob.

I hope this isn't going to be a problem on your end? Please let me
know if any questions or concerns,

Mark