cmd/juju: destroy-service returns 0 when service is not in a state to respond to the destroy request
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
juju-core |
Triaged
|
High
|
Unassigned |
Bug Description
ubuntu@winton-02:~$ juju status mysql
environment: manual
machines:
"1":
agent-state: started
agent-version: 1.17.8.9001
dns-name: 10.245.67.7
instance-id: manual:10.245.67.7
series: trusty
hardware: arch=ppc64 cpu-cores=1 mem=8165M
services:
mysql:
charm: local:trusty/
exposed: false
life: dying
relations:
cluster:
- mysql
units:
mysql/0:
life: dying
machine: "1"
ubuntu@winton-02:~$ juju destroy-service mysql ; echo $?
0
juju destroy-service should return none zero as the unit mysql/0 is in an error state so the destroy-service action cannot proceed.
Changed in juju-core: | |
status: | New → Triaged |
importance: | Undecided → High |
milestone: | none → 2.0 |
tags: | added: destroy-environment hooks |
Disagree; the request to destroy the service has been accepted and acted upon. But it is reasonable (and kinder to the user) to change the behaviour of the associated units/machines, as in lp:1219902, of which I am marking this bug a dupe.