juju 29.-rc8 breaks manual provider spaces

[Michał Ajduk]

I've bootstrapped juju manual provider on a machine with 3 interfaces:
root@juju1:~# ip -4 a s
1: lo: <LOOPBACK,UP,LOWER_UP> mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000
    inet 127.0.0.1/8 scope host lo
       valid_lft forever preferred_lft forever
2: ens3: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1422 qdisc fq_codel state UP group default qlen 1000
    inet 172.27.90.240/23 brd 172.27.91.255 scope global dynamic ens3
       valid_lft 85966sec preferred_lft 85966sec
3: ens4: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1372 qdisc fq_codel state UP group default qlen 1000
    inet 10.0.0.22/24 brd 10.0.0.255 scope global dynamic ens4
       valid_lft 85966sec preferred_lft 85966sec
4: ens5: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1372 qdisc fq_codel state UP group default qlen 1000
    inet 10.1.0.187/24 brd 10.1.0.255 scope global dynamic ens5
       valid_lft 85966sec preferred_lft 85966sec

Juju bootstrapped, but I'm not able to get spaces to work:
osm@osm:~/osm-demo$ juju spaces
Name Space ID Subnets
alpha 0

osm@osm:~/osm-demo$ juju add-subnet 10.0.0.0/24 oam
ERROR cannot add subnet: cannot get provider subnets: subnets not supported

osm@osm:~/osm-demo$ juju subnets
No subnets to display.

Juju status for the controller also does not work:
osm@osm:~/osm-demo$ juju show-machine -m controller 0
{}
ERROR cannot obtain space information: CIDR "172.27.90.240/23" not valid

osm@osm:~/osm-demo$ juju status -m controller
ERROR cannot obtain space information: CIDR "172.27.90.240/23" not valid
osm@osm:~/osm-demo$ juju status
Model Controller Cloud/Region Version SLA Timestamp
default osm-controller osm_manual_cloud/default 2.9-rc8 unsupported 14:03:02Z

Model "admin/default" is empty.

Relevant log entries:
2021-03-31 13:46:10 INFO juju.worker.machiner machiner.go:149 setting addresses for "machine-0" to [local-machine:127.0.0.1 local-cloud:172.27.90.240 local-cloud:10.0.0.22 local-cloud:10.1.0.187 local-machine:::1]
2021-03-31 13:46:10 INFO juju.state machine.go:1770 machine "0" preferred private address changed from "public:172.27.90.240" to "public:172.27.90.240"
2021-03-31 13:46:10 INFO juju.state machine.go:1778 machine "0" preferred public address changed from "public:172.27.90.240" to "public:172.27.90.240"
2021-03-31 13:46:10 INFO juju.apiserver.common.networkingcommon linklayer.go:169 processing link-layer devices for machine "0"
2021-03-31 13:46:10 INFO juju.apiserver.common.networkingcommon networkconfigapi.go:367 machine "0": adding new device "lo" () with addresses [127.0.0.1/8 ::1/128]
2021-03-31 13:46:10 INFO juju.apiserver.common.networkingcommon networkconfigapi.go:367 machine "0": adding new device "ens3" (fa:16:3e:ab:06:6b) with addresses [172.27.90.240/23]
2021-03-31 13:46:10 INFO juju.apiserver.common.networkingcommon networkconfigapi.go:367 machine "0": adding new device "ens4" (fa:16:3e:6b:1e:87) with addresses [10.0.0.22/24]
2021-03-31 13:46:10 INFO juju.apiserver.common.networkingcommon networkconfigapi.go:367 machine "0": adding new device "ens5" (fa:16:3e:3d:7b:10) with addresses [10.1.0.187/24]
2021-03-31 13:46:11 ERROR juju.worker.dependency engine.go:671 "machiner" manifold worker returned unexpected error: setting machine addresses: CIDR "172.27.90.240/23" not valid
2021-03-31 13:46:38 ERROR juju.worker.dependency engine.go:671 "peer-grouper" manifold worker returned unexpected error: computing desired peer group: updating member addresses: juju-ha-space is not set and these nodes have more than one usable address: 0

When trying to bootstrap with just a single interface, it also fails the same way, relevant logs:
2021-03-31 14:22:34 INFO juju.container.lxd manager.go:71 Availability zone will be empty for this container manager
2021-03-31 14:22:37 ERROR juju.worker.dependency engine.go:671 "firewaller" manifold worker returned unexpected error: CIDR "172.27.91.84/23" not valid

[Achilleas Anagnostopoulos (achilleasa)]

PR https://github.com/juju/juju/pull/12845 includes a fix for 2.9