Howdy,

Blake J pointed us (UAB Research Computing) to your nice slurm-wrapper repo :-)

While job-history and system-busy work great on our cluster, nodes-busy is crashing on our cluster with the following:

❯ ./bin/nodes-busy

nodes-busy: visualize live system resource usage.
Trouble seeing the output? Try 'nodes-busy --ascii'

Traceback (most recent call last):
  File "./bin/nodes-busy", line 766, in <module>
    merged = merge(job_data, node_data)
  File "./bin/nodes-busy", line 330, in merge
    nodes_dictionary[node]["JOBS"][job] = {"CPUs":cpus, "GPUs":jobs_dictionary[job]["GPUAlloc"],"EndTime":jobs_dictionary[job]["EndTime"],"Partition":jobs_dictionary[job]["Partition"],"Restarts":jobs_dictionary[job]["Restarts"]}
KeyError: 'GPUAlloc'

I'm digging through the code, but figured I'd post this for tracking purposes.

Thanks, Mike

Howdy, noticed the following error today when running nodes-busy on our cluster.

> nodes-busy

Traceback (most recent call last):
  File "/share/apps/rc/bin/nodes-busy", line 1177, in <module>
    job_data = get_scontrol_job_data()
  File "/share/apps/rc/bin/nodes-busy", line 370, in get_scontrol_job_data
    output = out.decode('utf-8').split("\n")
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xad in position 53196: invalid start byte

I have no idea what's in position 53196, but I fixed the issue with the following addition to output.decode to ignore errors:

diff --git a/bin/nodes-busy b/bin/nodes-busy
index 6fb1744..aee728f 100755
--- a/bin/nodes-busy
+++ b/bin/nodes-busy
@@ -367,7 +367,7 @@ def get_scontrol_job_data(target_job = None):
         sys.exit(1)

     # Split up space-delimited output into a job dictionary
-    output = out.decode('utf-8').split("\n")
+    output = out.decode('utf-8', 'ignore').split("\n")
     for job in output:
         details = job.split(' ')
         for i in details:

Added new argument to past-jobs to support selecting jobs by someone other than the user running the script:

$ past-jobs -d 30 -u flakrat

                         Jobs submitted by user flakrat in last 30 days.

JobID    Start       User            JobName         Partition  Account    State       ExitCode
------------------------------------------------------------------------------------------------
14364561 2022-06-02  flakrat         hostname        express    flakrat    CANCELLED        0:0
14385956 2022-06-05  flakrat         hostname        express    flakrat    COMPLETED        0:0
14850816 2022-06-12  flakrat         hostname        express    flakrat    COMPLETED        0:0