juju info reports misleading information

[Felipe Reyes]

The OpenStack Charms publish charms for specific architecture (they are not multi-arch), so when running "juju info" without passing the "--arch" flag the output can be misleading for users not aware that charms can be arch-specific.

[Test Case]

$ uname -a
Linux kaytoo 6.2.0-31-generic #31-Ubuntu SMP PREEMPT_DYNAMIC Mon Aug 14 13:42:26 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
$ juju_3pt1 info octavia
...
  2023.1/stable: ee12597 2023-08-22 (190) 37MB arm64 ubuntu@22.04, ubuntu@22.10, ubuntu@23.04
  2023.1/candidate: ↑
  2023.1/beta: ↑
  2023.1/edge: ↑

Expected result:

The charm's revision(s) displayed is the same arch where the command was run.

Actual result:

The revision displayed is for arm64 and not amd64

[Joseph Phillips (manadart)]

Can you tell me:
- What is the architecture of the controller?
- Is there any configured default arch for the current model?
- Is octavia already deployed in the model?

[Felipe Reyes (freyes)]

On Thu, 2023-09-14 at 10:26 +0000, Joseph Phillips wrote:
> Can you tell me:
> - What is the architecture of the controller?

amd64 (controller and client)

> - Is there any configured default arch for the current model?

no.

$ juju model-constraints

$ # ^ emptry output.

> - Is octavia already deployed in the model?

no, the model is empty

$ juju status
Model Controller Cloud/Region Version SLA Timestamp
test lxd-3pt1 localhost/localhost 3.1.5 unsupported 08:59:40-03:00

Model "admin/test" is empty.

[Felipe Reyes (freyes)]

here I'm attaching the output of "charmcraft status octavia" for you get a more comprehensive list of the charm's revisions, architectures and bases this charm supports

[Harry Pidcock (hpidcock)]

Why it is showing arm64? probably because it is the highest revision arm64=190 amd64=187

Confirmed this is happening in 3.3 HEAD too.

[Felipe Reyes (freyes)]

Adding a bug task for the snapstore-server project, because we are seeing the same problem with the web view.

[Felipe Reyes (freyes)]

I understand this may be considered as low priority issue, although we have had many users confused with questions about why there is no charm for amd64, or why they got installed charm's revision 119 while they see there is a 121 revision published, etc. Can this bug be re prioritized to High?