Ticket 8023

Summary: Expose node-count as a % filename pattern
Product: Slurm Reporter: Carl Ponder <CPonder>
Component: AccountingAssignee: Jacob Jenson <jacob>
Status: RESOLVED INVALID QA Contact:
Severity: 6 - No support contract    
Priority: ---    
Version: 20.02.x   
Hardware: Linux   
OS: Linux   
Site: -Other- Slinky Site: ---
Alineos Sites: --- Atos/Eviden Sites: ---
Confidential Site: --- Coreweave sites: ---
Cray Sites: --- DS9 clusters: ---
Google sites: --- HPCnow Sites: ---
HPE Sites: --- IBM Sites: ---
NOAA SIte: --- NoveTech Sites: ---
Nvidia HWinf-CS Sites: --- OCF Sites: ---
Recursion Pharma Sites: --- SFW Sites: ---
SNIC sites: --- Tzag Elita Sites: ---
Linux Distro: --- Machine Name:
CLE Version: Version Fixed:
Target Release: --- DevPrio: ---
Emory-Cloud Sites: ---

Description Carl Ponder 2019-10-30 06:51:51 MDT 
I'm running a scaling study, with a sequence of jobs that vary by node count.
I'd like to be able to include the node counts in the names of the log files that I generate, i.e.

     #SBATCH -o %j.log.medium.%N_nodes

would generate log files with names like

     8635.log.4_nodes
     8359.log.8_nodes

etc. While you could argue that studies like this could vary any parameter, and how could you expose them all (?), but the node-count is controlled on the "sbatch" command-line (or batch control-file header) and should be available at the job startup just like the %j Job ID.