memory Limit is returning invalid result for /container/stats REST API on macOS

[benoitf]

Is this a BUG REPORT or FEATURE REQUEST? (leave only one on its own line)

/kind bug

Description

The field memory_stats.limit is returning an invalid value

Steps to reproduce the issue:

1. Start a dummy container like fedora

$ MY_DUMMY_CONTAINER=$(podman run -d fedora sleep 20000)

2. Display the memory of the container from inside the container

$ podman exec -ti $MY_DUMMY_CONTAINER cat /proc/meminfo | head -3
MemTotal:        2018452 kB
MemFree:          860084 kB
MemAvailable:    1556036 kB

3. Now, read the value returned by the REST API

curl --silent --unix-socket /Users/benoitf/.local/share/containers/podman/machine/podman-machine-default/podman.sock "http:/v1.41/containers/$MY_DUMMY_CONTAINER/stats?stream=false&one-shot=true"  | jq .memory_stats
{
  "usage": 94208,
  "max_usage": 18446744073709552000,
  "limit": 18446744073709552000
}

Describe the results you received:

  "limit": 18446744073709552000

it seems this value is like the maximum value of the int/long/etc

Describe the results you expected:

something displaying around 2GB / 2Gi because it's the size of the machine.

  "limit": 2066894848

and based from
https://docs.docker.com/engine/api/v1.41/#tag/Container/operation/ContainerStats

used_memory = memory_stats.usage - memory_stats.stats.cache
available_memory = memory_stats.limit
Memory usage % = (used_memory / available_memory) * 100.0

Additional information you deem important (e.g. issue happens only occasionally):

Output of podman version:

4.2.1

Output of podman info:

4.2.1

Package info (e.g. output of rpm -q podman or apt list podman):

(paste your output here)

Have you tested with the latest version of Podman and have you checked the Podman Troubleshooting Guide? (https://github.com/containers/podman/blob/main/troubleshooting.md)

Yes/No

Additional environment details (AWS, VirtualBox, physical, etc.):

[vrothberg]

Thanks for reaching out, @benoitf!

@giuseppe, can you take a look? This looks like a cgroup stats issue at first glance.

[giuseppe]

Docker caps the value returned from the cgroup to the maximum available memory, so I guess we are forced to do the same thing for compatibility, opened a PR: #15811