filtering juju status by machine id gives inconsistent results

[Andrea Ieri]

I often find myself running `juju status <id>` to obtain a view of every unit running on a specific machine. This works great, except that it sometimes returns units from unrelated machines. I have not been able to find a pattern to this, sometimes it's `juju status 0` that returns too much, sometimes it's `juju status 8` or some other id.

This can easily be reproduced by deploying basically any model and iterating through all machine ids: several will return too many entries.

For example, in a bionic-queens openstack-on-lxd deployment:

$ juju status 4
Model Controller Cloud/Region Version SLA Timestamp
aieri-openstack lxd localhost/localhost 2.7.0 unsupported 08:35:04Z

App Version Status Scale Charm Store Rev OS Notes
ceph-osd 12.2.12 active 1 ceph-osd jujucharms 294 ubuntu
openstack-dashboard 13.0.2 active 1 openstack-dashboard jujucharms 297 ubuntu

Unit Workload Agent Machine Public address Ports Message
ceph-osd/0* active idle 4 10.0.8.64 Unit is ready (1 OSD)
openstack-dashboard/0* active idle 21 10.0.8.53 80/tcp,443/tcp Unit is ready

Machine State DNS Inst id Series AZ Message
4 started 10.0.8.64 juju-c3090d-4 bionic Running
21 started 10.0.8.53 juju-c3090d-21 bionic Running

Here are some stats:

$ for i in `seq 0 23`; do juju status $i | grep -A1000 '^Machine' | grep -vE 'Machine|^$' | wc -l; done | sort | uniq -c
     17 1
      4 2
      2 3
      1 6

Basically, in my deployment filtering is correct for 17 out of 24 machines, but fails for the remaining 7, which return 2, 3, or even 6 entries.

Here is the worst one:

$ juju status 8
Model Controller Cloud/Region Version SLA Timestamp
aieri-openstack lxd localhost/localhost 2.7.0 unsupported 08:48:18Z

App Version Status Scale Charm Store Rev OS Notes
ceph-radosgw 12.2.12 active 1 ceph-radosgw jujucharms 283 ubuntu
cinder 12.0.7 active 1 cinder jujucharms 297 ubuntu
cinder-ceph 12.0.7 active 0 cinder-ceph jujucharms 251 ubuntu
gnocchi 4.2.5 active 1 gnocchi jujucharms 30 ubuntu
heat 10.0.2 active 1 heat jujucharms 271 ubuntu
nova-cloud-controller 17.0.11 active 1 nova-cloud-controller jujucharms 339 ubuntu
openstack-dashboard 13.0.2 active 1 openstack-dashboard jujucharms 297 ubuntu

Unit Workload Agent Machine Public address Ports Message
ceph-radosgw/0* active idle 7 10.0.8.37 80/tcp Unit is ready
cinder/0* active idle 8 10.0.8.70 8776/tcp Unit is ready
  cinder-ceph/0* active idle 10.0.8.70 Unit is ready
gnocchi/0* active idle 12 10.0.8.42 8041/tcp Unit is ready
heat/0* active idle 13 10.0.8.18 8000/tcp,8004/tcp Unit is ready
nova-cloud-controller/0* active idle 19 10.0.8.81 8774/tcp,8778/tcp Unit is ready
openstack-dashboard/0* active idle 21 10.0.8.53 80/tcp,443/tcp Unit is ready

Machine State DNS Inst id Series AZ Message
7 started 10.0.8.37 juju-c3090d-7 bionic Running
8 started 10.0.8.70 juju-c3090d-8 bionic Running
12 started 10.0.8.42 juju-c3090d-12 bionic Running
13 started 10.0.8.18 juju-c3090d-13 bionic Running
19 started 10.0.8.81 juju-c3090d-19 bionic Running
21 started 10.0.8.53 juju-c3090d-21 bionic Running

[Andrea Ieri (aieri)]

...sorry for the mangled output. Here it is again in a pastebin: https://pastebin.canonical.com/p/PWfnrkxjV9/

[Junien F (axino)]

Related (if not duplicate) : bug 1809411

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

This bug has not been updated in 2 years, so we're marking it Low importance. If you believe this is incorrect, please update the importance.

[Haw Loeung (hloeung)]

I think this is LP;1988709. `juju status 4` returns machine `21` because it has `443/tcp`. `juju status 8` because `80/tcp`, `8776/tcp`, `8041/tcp`, etc.