What were you trying to do that didn't work?

Inodes exhaustion happens when continuous idmapped mounts are used with successive podman runs.

What is the impact of this issue to you?

A business need that requires the use of dynamically created UID/GID mappings when running containers in a rootful environment. Unfortunately, this results in the exhaustion of inodes in the filesystem hosting the image store.

Please provide the package NVR for which the bug is seen:

podman-5.4.0-9.el9_6.x86_64.rpm

How reproducible is this bug?:

Fully

Steps to reproduce

[root@localhost overlay]# podman run -dt --uidmap 0:2331011011:65536 --gidmap 0:2331011011:65536 --uidmap 70000011:19999999:1000 ubi9-minimal
4f901e124e029dc7ae74d66ef1a298d947a8a94eef9aca247e1ec024779233a5

[root@localhost overlay]# df -i /var/lib/containers/storage/overlay
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/mapper/rhel-root 18814976 283589 18531387 2% /var/lib/containers/storage/overlay

[root@localhost overlay]# podman run -dt --uidmap 0:233101109:65536 --gidmap 0:233101109:65536 --uidmap 700000111:19999999:1000 ubi9-minimal
273614b825a72303f6aa7e8e05a8b793e999d2853226f4d61425cd7622fa0bb0

[root@localhost overlay]# df -i /var/lib/containers/storage/overlay
Filesystem Inodes IUsed IFree IUse% Mounted on
/dev/mapper/rhel-root 18814976 286295 18528681 2% /var/lib/containers/storage/overlay

Consumption of almost 3000 inodes for the same image.

Additional Information :

https://github.com/containers/podman/discussions/26205