oblachno-oss/grm
3
0
Fork 0
Code Issues 20 Pull Requests Actions Packages Projects Releases 14 Wiki Activity

GRM-159: fix docker-prune.timer schedule and gitea-runner.service StartLimit placement #244

Merged
emo merged 1 commits from GRM-159-fix-runner-timer-and-service-template into master 2026-08-06 09:23:32 +00:00
Conversation 3 Commits 1 Files Changed 5
+42 -10
1 Commits
Author SHA1 Message Date
Emil SimeonovandDevin <158243242+devin-ai-integration[bot]@users.noreply.github.com> 6a0eef680d fix: move StartLimit to [Unit] and make prune timer reload conditional
CI / validate (pull_request) Successful in 1m34s
Details
CI / molecule-tests (2) (pull_request) Successful in 4m1s
Details
CI / molecule-tests (1) (pull_request) Successful in 4m40s
Details
CI / molecule-tests (4) (pull_request) Successful in 5m35s
Details
CI / molecule-tests (3) (pull_request) Successful in 6m14s
Details
CI / molecule-tests (6) (pull_request) Successful in 6m40s
Details
CI / molecule-tests (5) (pull_request) Successful in 8m27s
Details
CI / auto-merge (pull_request) Successful in 1m3s
Details 
Two issues found on all deployed runners:

1. gitea-runner.service had StartLimitIntervalSec and StartLimitBurst
   in [Service] section — systemd ignores them with "Unknown key name"
   warning. Moved to [Unit] section where they belong.

2. docker-prune.timer and gitea-runner.service templates were deployed
   but systemd daemon-reload ran unconditionally (even when the file
   didn't change). Added register+conditional so daemon-reload and
   service restart only happen when the template actually changes.

The deployed runners have a broken docker-prune.timer with invalid
OnCalendar schedule (*-* * */6:00:00 instead of *-*-* 00/6:00:00)
because an older version of the role was used. Reinstalling with the
fixed role will correct this.

Co-Authored-By: Devin <158243242+devin-ai-integration[bot]@users.noreply.github.com>
 2026-08-06 11:12:07 +02:00