laforge submitted this change.

View Change

ansible: obsworker: restart on failure

The obsworker service can fail if it can't reach the OBS server after it
started. Let it retry continuously until the server comes up again.

As the original service is of type "oneshot", Restart=on-failure must be
used instead of Restart=always (https://unix.stackexchange.com/a/777329).

Change-Id: I81fe02984d6c227a482b0b2421a967def1e208d5
---
A ansible/roles/osmocom-obs-worker/files/obsworker.conf
M ansible/roles/osmocom-obs-worker/tasks/obs.yml
2 files changed, 12 insertions(+), 0 deletions(-)

diff --git a/ansible/roles/osmocom-obs-worker/files/obsworker.conf b/ansible/roles/osmocom-obs-worker/files/obsworker.conf
new file mode 100644
index 0000000..cb9a4dd
--- /dev/null
+++ b/ansible/roles/osmocom-obs-worker/files/obsworker.conf
@@ -0,0 +1,3 @@
+[Service]
+Restart=on-failure
+RestartSec=10
diff --git a/ansible/roles/osmocom-obs-worker/tasks/obs.yml b/ansible/roles/osmocom-obs-worker/tasks/obs.yml
index 48024ec..f13195a 100644
--- a/ansible/roles/osmocom-obs-worker/tasks/obs.yml
+++ b/ansible/roles/osmocom-obs-worker/tasks/obs.yml
@@ -58,6 +58,15 @@
     name: mount-securityfs.service
     enabled: yes
 
+- name: create /etc/systemd/system/obsworker.service.d/
+  file:
+    path: /etc/systemd/system/obsworker.service.d/
+
+- name: configure systemd service to restart on failure
+  ansible.builtin.copy:
+    src: obsworker.conf
+    dest: /etc/systemd/system/obsworker.service.d/override.conf
+
 - name: ensure the systemd service is installed
   systemd:
     name: obsworker.service

To view, visit change 43206. To unsubscribe, or for help writing mail filters, visit settings.