/tmp is fills up on agent nodes due to bad job-discipline. Setting up Jenkins-level jobs to clean /tmp will fail once Jenkins declares /tmp over-full and offlines the node (which blocks the Jenkins-level jobs from attempting further cleaning attempts)
Problem is avoided if using Agents via ASGs that are configured for frequent rebuilds.
Expected Behavior:
The /tmp partition should not fill and, being a "temporary" storage location, likely that nothing in /tmp should be any older than a few days.
Actual Behavior:
/tmp fills till the service offlines
Fix recommendation:
Add a daily cron job to clean up after Jenkins makes a mess of things.
ferricoxide
commented
6 years ago
Problem Description:
/tmpis fills up on agent nodes due to bad job-discipline. Setting up Jenkins-level jobs to clean/tmpwill fail once Jenkins declares/tmpover-full and offlines the node (which blocks the Jenkins-level jobs from attempting further cleaning attempts)Problem is avoided if using Agents via ASGs that are configured for frequent rebuilds.
Expected Behavior:
The
/tmppartition should not fill and, being a "temporary" storage location, likely that nothing in/tmpshould be any older than a few days.Actual Behavior:
/tmpfills till the service offlinesFix recommendation:
Add a daily cron job to clean up after Jenkins makes a mess of things.