While running a load test for COS Lite, a workloadless charm received a StopContainer (haven't figured out why yet).
As a result, a `stop` hooked was followed by a `start` hook with nothing in between and nothing after.
$ juju debug-log -i unit-scrape-target-0 --replay | grep Emitting unit-scrape-target-0: 06:56:26 DEBUG unit.scrape-target/0.juju-log Emitting Juju event update_status. unit-scrape-target-0: 07:00:37 DEBUG unit.scrape-target/0.juju-log Emitting Juju event stop. unit-scrape-target-0: 07:00:44 DEBUG unit.scrape-target/0.juju-log Emitting Juju event start. unit-scrape-target-0: 07:05:36 DEBUG unit.scrape-target/0.juju-log Emitting Juju event update_status. unit-scrape-target-0: 07:09:56 DEBUG unit.scrape-target/0.juju-log Emitting Juju event update_status.
$ juju show-status-log scrape-target/0 Time Type Status Message 06 Apr 2023 11:55:41-04:00 juju-unit allocating 06 Apr 2023 11:55:41-04:00 workload waiting installing agent 06 Apr 2023 11:56:26-04:00 workload waiting agent initializing 06 Apr 2023 11:56:30-04:00 workload maintenance installing charm software 06 Apr 2023 11:56:30-04:00 juju-unit executing running install hook 06 Apr 2023 11:56:32-04:00 juju-unit executing running metrics-endpoint-relation-created hook 06 Apr 2023 11:56:34-04:00 juju-unit executing running leader-elected hook 06 Apr 2023 11:56:34-04:00 juju-unit executing running config-changed hook 06 Apr 2023 11:56:36-04:00 juju-unit executing running start hook 06 Apr 2023 11:56:37-04:00 juju-unit executing running metrics-endpoint-relation-joined hook for scrape-config/0 06 Apr 2023 11:56:38-04:00 juju-unit executing running metrics-endpoint-relation-changed hook for scrape-config/0 06 Apr 2023 11:57:16-04:00 juju-unit idle 06 Apr 2023 11:57:16-04:00 workload active 07 Apr 2023 07:00:36-04:00 workload maintenance stopping charm software 07 Apr 2023 07:00:36-04:00 juju-unit executing running stop hook 07 Apr 2023 07:00:43-04:00 juju-unit executing running start hook 07 Apr 2023 07:01:27-04:00 juju-unit idle 07 Apr 2023 07:01:27-04:00 workload maintenance
In contrast, when I kubectl delete the pod, I see the full upgrade sequence:
unit-scrape-target-0: 10:40:11 DEBUG unit.scrape-target/0.juju-log Emitting Juju event stop. unit-scrape-target-0: 10:40:19 DEBUG unit.scrape-target/0.juju-log Emitting Juju event upgrade_charm. unit-scrape-target-0: 10:40:19 DEBUG unit.scrape-target/0.juju-log Emitting Juju event config_changed. unit-scrape-target-0: 10:40:20 DEBUG unit.scrape-target/0.juju-log Emitting Juju event start.
While running a load test for COS Lite, a workloadless charm received a StopContainer (haven't figured out why yet).
As a result, a `stop` hooked was followed by a `start` hook with nothing in between and nothing after.
$ juju debug-log -i unit-scrape- target- 0 --replay | grep Emitting target- 0: 06:56:26 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event update_status. target- 0: 07:00:37 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event stop. target- 0: 07:00:44 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event start. target- 0: 07:05:36 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event update_status. target- 0: 07:09:56 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event update_status.
unit-scrape-
unit-scrape-
unit-scrape-
unit-scrape-
unit-scrape-
$ juju show-status-log scrape-target/0 endpoint- relation- created hook endpoint- relation- joined hook for scrape-config/0 endpoint- relation- changed hook for scrape-config/0
Time Type Status Message
06 Apr 2023 11:55:41-04:00 juju-unit allocating
06 Apr 2023 11:55:41-04:00 workload waiting installing agent
06 Apr 2023 11:56:26-04:00 workload waiting agent initializing
06 Apr 2023 11:56:30-04:00 workload maintenance installing charm software
06 Apr 2023 11:56:30-04:00 juju-unit executing running install hook
06 Apr 2023 11:56:32-04:00 juju-unit executing running metrics-
06 Apr 2023 11:56:34-04:00 juju-unit executing running leader-elected hook
06 Apr 2023 11:56:34-04:00 juju-unit executing running config-changed hook
06 Apr 2023 11:56:36-04:00 juju-unit executing running start hook
06 Apr 2023 11:56:37-04:00 juju-unit executing running metrics-
06 Apr 2023 11:56:38-04:00 juju-unit executing running metrics-
06 Apr 2023 11:57:16-04:00 juju-unit idle
06 Apr 2023 11:57:16-04:00 workload active
07 Apr 2023 07:00:36-04:00 workload maintenance stopping charm software
07 Apr 2023 07:00:36-04:00 juju-unit executing running stop hook
07 Apr 2023 07:00:43-04:00 juju-unit executing running start hook
07 Apr 2023 07:01:27-04:00 juju-unit idle
07 Apr 2023 07:01:27-04:00 workload maintenance
In contrast, when I kubectl delete the pod, I see the full upgrade sequence:
unit-scrape- target- 0: 10:40:11 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event stop. target- 0: 10:40:19 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event upgrade_charm. target- 0: 10:40:19 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event config_changed. target- 0: 10:40:20 DEBUG unit.scrape- target/ 0.juju- log Emitting Juju event start.
unit-scrape-
unit-scrape-
unit-scrape-