[HTCondor-users] Question about POST scripts in a DAG workflow

[email.nsimpson@xxxxxxxxx]

Hi all,

I have a nested DAG setup as follows:

SUBDAG EXTERNAL dag1 dag1.dag
SCRIPT POST dag1 script.py arg1 arg2 arg3
â (repeats with other dags)

where arg1-3 are a set of command-line args I feed to script.py.

I wrote script.py to scrape the dag log files from dag1 and send an automated email. This file runs fine locally, and even is fed the absolute path to the log files as an argument to make sure the script can find them. However, it doesnât work in practice, and I canât see the output or errors from the POST script anywhere in the dagman log files.

My questions are:
- Is this the right way to feed args to a POST script?
- Are POST scripts run locally (submit machine) or on a node?

And also, my current dagman out is below for convenience, maybe this has an explanation?

12/04/20 00:38:41 Running POST script of Node dag1...
12/04/20 00:38:41 Warning: mysin has length 0 (ignore if produced by DAGMan; see gittrac #4987, #5031)
12/04/20 00:38:41 Create_Process(script.py): child failed with errno 8 (Exec format error) before exec()
12/04/20 00:38:41   error: daemonCore->Create_Process() failed; POST script of Job dag1 failed
12/04/20 00:38:41 Initializing user log writer for ./dag.dag.nodes.log, (1064990.0.0)
12/04/20 00:38:41 Event: ULOG_POST_SCRIPT_TERMINATED for HTCondor Node iteration00 (1064990.0.0) {12/04/20 00:38:41}
12/04/20 00:38:41 POST Script of node dag1 failed with status 1
12/04/20 00:38:41 POST for Node dag1 returned 1

Best, Nathan