I tried running this on my laptop and was surpriced it waned reboot. That was due creating dummy user. That should not cause reboot. Probably none of these really:

builder_packages is changed or
        package_update is changed or
        builder_socket is changed or
        builder_user is changed

perhaps just moving (re)start of osbuild-composer.socket osbuild-composer.service osbuild-local-worker.socket to handler if install/update event happened.

I am building an RHEL 8.7 image on an AWS build host. After the first successful run, if I want to build an image again, I get this error:

TASK [Push image blueprint to osbuilder] ***************************************
task path: /root/.ansible/pull/ip-.ec2.internal/build_image.yml:13
fatal: [localhost]: FAILED! => {"changed": true, "cmd": ["composer-cli", "blueprints", "push", "base-image.toml"], "delta": "0:00:01.243122", "end": "2023-04-07", "msg": "non-zero return code", "rc": 1, "start": "2023-04-07*", "stderr": "ERROR: Push TOML: Post "http://localhost/api/v1/blueprints/new\": read unix @->/run/weldr/api.socket: read: connection reset by peer", "stderr_lines": ["ERROR: Push TOML: Post "http://localhost/api/v1/blueprints/new": read unix @->/run/weldr/api.socket: read: connection reset by peer"],"stdout": "", "stdout_lines": []}

Well, if I run these commands below or reboot the ec2 instance, I can successfully create another image again.

systemctl stop osbuild*

systemctl reset-failed osbuild*

I wonder why it's happening because I can see that builder_uninstall.yml is cleaning up the previous image-builder stuff.
It's not a big issue, as I can still get the job done but would be nice if I could take care of it programmatically. I have tried several things but nothing seems to work.

I changed the below parameters in the playbook and used ansible pull to run the playbook but the image does not get created in /tmp/images:
rhel_image_git_remote_repo: https://gitlab.com/myblueprints.git
rhel_image_output_type: ami

There is no error in the ansible run

I started having this issue recently, using the latest code from this repo. My blueprint file is also pretty basic. Not sure what's going to resolve this.

TASK [Download image from osbuilder] *******************************************
fatal: [localhost]: FAILED! => {"changed": true, "cmd": ["composer-cli", "compose", "image", "4c15ab13-f48d-4ab2-a1b5-5c33f59b9a82"], "delta": "0:00:00.024304", "end": "2023-05-16 14:17:29.163723", "msg": "non-zero return code", "rc": 1, "start": "2023-05-16 14:17:29.139419", "stderr": "ERROR: BuildInWrongState: Build 4c15ab13-f48d-4ab2-a1b5-5c33f59b9a82 is in wrong state: RUNNING", "stderr_lines": ["ERROR: BuildInWrongState: Build 4c15ab13-f48d-4ab2-a1b5-5c33f59b9a82 is in wrong state: RUNNING"], "stdout": "", "stdout_lines": []}

PLAY RECAP *****************************************************************

Blueprint file:

Base definition

name = "base-image"
description = "Base Image"
version = "2023.03.13"
distro = "rhel-87"

Boot parameters

[customizations.kernel]
append = "no_timer_check console=hvc0 LANG=en_US.UTF-8 console=ttyS0,115200n8 elevator=noop net.ifnames=0 biosdevname=0"

Filesystem

[[customizations.filesystem]]
mountpoint = "/boot"
size = "1 GiB"
[[customizations.filesystem]]
mountpoint = "/"
size = "8 GiB"
[[customizations.filesystem]]
mountpoint = "/home"
size = "2 GiB"
[[customizations.filesystem]]
mountpoint = "/tmp"
size = "2 GiB"
[[customizations.filesystem]]
mountpoint = "/var"
size = "8 GiB"
[[customizations.filesystem]]
mountpoint = "/var/log"
size = "2 GiB"
[[customizations.filesystem]]
mountpoint = "/var/log/audit"
size = "2 GiB"
[[customizations.filesystem]]
mountpoint = "/var/tmp"
size = "2 GiB"

Locale

Timezone

[customizations.timezone]
timezone = "America/New_York"

Groups

[[groups]]
name="Minimal Install"

[[customizations.user]]
name = "ec2-user"
groups = ["wheel"]