SymbX/systemd
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

job: Don't discard propagated restart jobs when unit is activating

Browse Source 
When a service unit Requires= a socket, and the socket is restarted while the
service is in state=activating, the propagated restart is being discarded. This
is contrary to the documentation for Requires=, which states "this unit will be
stopped (or restarted) if one of the other units is explicitly stopped (or
restarted)".
This commit is contained in:
Andrew Stone
2022-08-25 16:38:23 -07:00
parent 01532c2d16
commit dc06321fe3
4 changed files with 43 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files
test/units/testsuite-03.sh
+9
View File
@@ -50,6 +50,15 @@ systemctl stop hello.service sleep.service hello-after-sleep.target
# TODO: add more job queueing/merging tests here.
# Test that restart propagates to activating units
systemctl -T --no-block start always-activating.service
systemctl list-jobs | grep 'always-activating.service'
ACTIVATING_ID_PRE=$(systemctl show -P InvocationID always-activating.service)
systemctl -T start always-activating.socket # Wait for the socket to come up
systemctl -T restart always-activating.socket
ACTIVATING_ID_POST=$(systemctl show -P InvocationID always-activating.service)
[ "$ACTIVATING_ID_PRE" != "$ACTIVATING_ID_POST" ] || exit 1
# Test for irreversible jobs
systemctl start unstoppable.service