Allocated Ports are duplicated on Alloc info

[cgbaker]

Nomad version

Nomad v1.0.0-beta3 (fcb32ef)

Issue

the ports listed under AllocatedResources contain a growing list of duplicated allocations

Reproduction steps

1. create a job with a static port resource (I used the traefik job)
2. note that on every query to the allocation, the AllocatedResources includes an increasing number of copies of the port reservations:

$ nomad alloc status -json 44302903  | jq '.AllocatedResources.Shared.Ports | length'
56

$ nomad alloc status -json 44302903  | jq '.AllocatedResources.Shared.Ports | length'
58

I suspect that the issue is that the allocGet endpoint is calling canonicalize on the in-mem structs object:

nomad/command/agent/alloc_endpoint.go

Line 130 in 974039b

alloc.AllocatedResources.Canonicalize()

[cgbaker]

still an issue on 0.12.8
also, PR #9368 potentially does a double-copy (I missed that there is a potential alloc copy in there if the job payload is decoded)

[tgross]

@cgbaker should this have been closed by #9367 or is this something we need to get fixed before 1.0-rc?

[tgross]

Checked with current master:

$ nomad job run ./traefik.nomad
Job Warnings:
1 warning(s):

* Group "traefik" has warnings: 1 error occurred:
        * 1 error occurred:
        * Task "traefik": task network resources have been deprecated as of Nomad 0.12.0. Please configure networking via group network block.





==> Monitoring evaluation "75abe09a"
    Evaluation triggered by job "traefik"
    Allocation "a223dff5" created: node "ae49493f", group "traefik"
==> Monitoring evaluation "75abe09a"
    Evaluation within deployment: "5371c3e0"
    Evaluation status changed: "pending" -> "complete"
==> Evaluation "75abe09a" finished with status "complete"

$ nomad alloc status -json a223dff5 | jq '.AllocatedResources.Shared.Ports | length'
2
$ nomad alloc status -json a223dff5 | jq '.AllocatedResources.Shared.Ports | length'
2

Looks like we're good?

[tgross]

Verified with RC1.

[github-actions]

I'm going to lock this issue because it has been closed for 120 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.