[manual-machines] ERROR juju-ha-space is not set and a unique usable address was not found for machines: 0
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Fix Released
|
High
|
Joseph Phillips |
Bug Description
Hello,
When bootstrapping manual machines on the same subnet it is not possible to enable HA despite all controllers belonging to the same subnet and space within juju.
[Reproducer]
1. Create three VMs on the same subnet, for example /24.
2. Bootstrap on one VM using manual cloud and add the other two VMs
3. juju enable-ha --to=1,2
ERROR juju-ha-space is not set and a unique usable address was not found for machines: 0
run "juju controller-config juju-ha-
Attempting to set juju-ha-space to alpha does not appear to work either:
root@juju-client:~# juju controller-config juju-ha-space=alpha
ERROR invalid config "juju-ha-
[Additional information]
root@juju-client:~# juju status -m controller
Model Controller Cloud/Region Version SLA Timestamp
controller juju-manual-default juju-manual/default 2.9.34 unsupported 14:11:50Z
Machine State Address Inst id Series AZ Message
0 started 10.66.240.6 manual: focal
1 down 10.66.240.7 manual:10.66.240.7 focal
2 down 10.66.240.222 manual:
root@juju-client:~# juju show-machine -m controller 0
model: controller
machines:
"0":
juju-status:
current: started
since: 24 Sep 2022 14:07:24Z
version: 2.9.34
hostname: juju-ctl-manual-1
dns-name: 10.66.240.6
ip-addresses:
- 10.66.240.6
- fd42:9a92:
- fd42:9a92:
instance-id: 'manual:'
machine-status:
current: pending
since: 24 Sep 2022 13:49:27Z
modificatio
current: idle
since: 24 Sep 2022 13:49:27Z
series: focal
network-
enp5s0:
- 10.66.240.6
- fd42:9a92:
- fd42:9a92:
gateway: 10.66.240.1 10.66.240.1 10.66.240.1
space: alpha
is-up: true
constraints: arch=amd64 mem=3584M
hardware: arch=amd64 cores=2 mem=15627M
controller-
root@juju-client:~# juju show-machine -m controller 1
model: controller
machines:
"1":
juju-status:
current: started
since: 24 Sep 2022 13:52:29Z
version: 2.9.34
hostname: juju-ctl-manual-2
dns-name: 10.66.240.7
ip-addresses:
- 10.66.240.7
- fd42:9a92:
- fd42:9a92:
instance-id: manual:10.66.240.7
machine-status:
current: pending
since: 24 Sep 2022 13:51:22Z
modificatio
current: idle
since: 24 Sep 2022 13:51:22Z
series: focal
network-
enp5s0:
- 10.66.240.7
- fd42:9a92:
- fd42:9a92:
gateway: 10.66.240.1 10.66.240.1 10.66.240.1
space: alpha
is-up: true
hardware: arch=amd64 cores=2 mem=15627M
root@juju-client:~# juju show-machine -m controller 2
model: controller
machines:
"2":
juju-status:
current: started
since: 24 Sep 2022 13:55:18Z
version: 2.9.34
hostname: juju-ctl-manual-0
dns-name: 10.66.240.222
ip-addresses:
- 10.66.240.222
- fd42:9a92:
- fd42:9a92:
instance-id: manual:
machine-status:
current: pending
since: 24 Sep 2022 13:54:24Z
modificatio
current: idle
since: 24 Sep 2022 13:54:24Z
series: focal
network-
enp5s0:
- 10.66.240.222
- fd42:9a92:
- fd42:9a92:
gateway: 10.66.240.1 10.66.240.1 10.66.240.1
space: alpha
is-up: true
hardware: arch=amd64 cores=2 mem=15627M
r
[Workaround]
None at this time,
Thanks,
Peter
description: | updated |
description: | updated |
description: | updated |
Changed in juju: | |
milestone: | none → 3.3.4 |
status: | In Progress → Fix Committed |
Changed in juju: | |
status: | Fix Committed → Fix Released |
Can you get me the output of `juju-spaces` in the controller model after you have added the machines?
The error might be misleading regarding machine 0 having no addresses in the space.
How it works is that Juju filters the machine addresses based on a local-cloud scope. If this yields a single address, we proceed. If not we require the operator declare which address to use via spaces.
So if you're just using "alpha" and it happens not to isolate a unique address, we can't proceed.
Does it work if you put the IPv4 subnet alone into a new space, say "beta" and use that as the HA space?