Machine is still marked 'down' after it comes back up

[Aaron Bentley]

I can force a machine to be treated as 'down' by running
juju run --machine 1 "sudo iptables-save > \$HOME/iptables; echo \"iptables-restore < \$HOME/iptables\"| sudo at now + 6 minutes; echo \"iptables --flush; iptables -P FORWARD DROP; iptables -P OUTPUT DROP; iptables -P INPUT DROP\"| sudo at now + 1 minutes"

Then, once juju notices that the machine is down, I can run:
$ juju run --machine 1 "echo I\'m back!" && juju status --format 'yaml'
I'm back!
model:
  name: abtest
  controller: abtest
  cloud: localhost
  region: localhost
  version: 2.1.1
...
  "1":
    juju-status:
      current: down
      message: agent is not communicating with the server
      since: 20 Mar 2017 10:49:03-04:00
      version: 2.1.1

The run operation blocks until the machine comes back up. After it succeeds, the machine is clearly not down.

Juju is cautious about marking a machine as down, and that makes sense. But any successful juju operation, "run" included, is proof that a machine is not down, and its status should be updated accordingly.

This issue was discovered while crafting a CI test. It will make CI harder to implement, because we cannot assume that an otherwise-functional machine is not marked "down".

[Canonical Juju QA Bot (juju-qa-bot)]

This bug has not been updated in 5 years, so we're marking it Expired. If you believe this is incorrect, please update the status.