CI: Run Nvidia workflow on UC22 #217

locnnil · 2025-01-20T16:19:20Z

Currently, the OS used on the Nvidia tests is hardcoded:

docker-snap/.github/workflows/testflinger/nvidia-job.yaml

distro: "noble"

However, we already have mechanisms in place to support testing on UC22 as well:

docker-snap/.github/workflows/testflinger/scripts/setup.sh

Lines 49 to 63 in 7d5e826

    
           setup_core() { 
        
             sudo snap install nvidia-core22 
        
             sudo snap install nvidia-assemble --channel 22/stable 
        
           } 
        
           setup() { 
        
             . /etc/os-release 
        
             install_docker 
        
             if [[ $ID == "ubuntu" ]]; then 
        
               setup_classic 
        
             elif [[ $ID == "ubuntu-core" ]]; then 
        
               setup_core

This PR updates the workflow to include tests for UC22 alongside the existing tests for Noble.

[skip ci]

.github/workflows/nvidia-test.yml

.github/workflows/testflinger/nvidia-job.yaml

.github/workflows/nvidia-test.yml

[skip ci] Signed-off-by: Lincoln Wallace <[email protected]>

Signed-off-by: Lincoln Wallace <[email protected]>

farshidtz · 2025-01-24T16:43:48Z

I've done some manual testing and there are a few other issues causing failure on Ubuntu Core 22, such as:

kernel and core snaps being refreshed and device rebooted after first boot, during the test setup
docker daemon not ready right after reboot
docker run command ran interactively

@locnnil I will commit fixes directly to your branch.

* Wait for snap changes * Wait for docker daemon after reboot * Run docker commands with sudo * Docker run in non-interactive mode

farshidtz · 2025-01-24T18:31:32Z

I still get this error randomely, on every 2-3 runs:

+ sudo snap install docker --channel=latest/edge
error: cannot perform the following tasks:
- Run install hook of "docker" snap if present (run hook "install": cannot perform operation: mount -t tmpfs /tmp/snap.rootfs_8C1znV: Permission denied)

Trying directly on the failing device gives the same result:

$ snap changes
ID   Status  Spawn               Ready               Summary
1    Done    today at 15:51 UTC  today at 15:51 UTC  Initialize system state
2    Done    today at 15:51 UTC  today at 15:51 UTC  Initialize device
3    Done    today at 15:51 UTC  today at 15:54 UTC  Auto-refresh snaps "core22", "pc-kernel"
4    Done    today at 15:54 UTC  today at 15:54 UTC  Hold auto-refreshes for all snaps
5    Error   today at 15:54 UTC  today at 15:55 UTC  Install "docker" snap from "latest/edge" channel

$ snap change 5
Status  Spawn               Ready               Summary
Done    today at 15:54 UTC  today at 15:55 UTC  Ensure prerequisites for "docker" are available
Undone  today at 15:54 UTC  today at 15:55 UTC  Download snap "docker" (3051) from channel "latest/edge"
Done    today at 15:54 UTC  today at 15:55 UTC  Fetch and check assertions for snap "docker" (3051)
Undone  today at 15:54 UTC  today at 15:55 UTC  Mount snap "docker" (3051)
Undone  today at 15:54 UTC  today at 15:55 UTC  Copy snap "docker" data
Undone  today at 15:54 UTC  today at 15:55 UTC  Setup snap "docker" (3051) security profiles
Undone  today at 15:54 UTC  today at 15:55 UTC  Make snap "docker" (3051) available to the system
Undone  today at 15:54 UTC  today at 15:55 UTC  Automatically connect eligible plugs and slots of snap "docker"
Undone  today at 15:54 UTC  today at 15:55 UTC  Set automatic aliases for snap "docker"
Undone  today at 15:54 UTC  today at 15:55 UTC  Setup snap "docker" aliases
Error   today at 15:54 UTC  today at 15:55 UTC  Run install hook of "docker" snap if present
Hold    today at 15:54 UTC  today at 15:55 UTC  Start snap "docker" (3051) services
Hold    today at 15:54 UTC  today at 15:55 UTC  Run configure hook of "docker" snap if present
Hold    today at 15:54 UTC  today at 15:55 UTC  Run health check of "docker" snap
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:network to snapd:network
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:support to snapd:docker-support
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:firewall-control to snapd:firewall-control
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:network-bind to snapd:network-bind
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:opengl to snapd:opengl
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:network-control to snapd:network-control
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:privileged to snapd:docker-support
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:docker-cli to docker:docker-daemon
Undone  today at 15:55 UTC  today at 15:55 UTC  Connect docker:home to snapd:home
Undone  today at 15:55 UTC  today at 15:55 UTC  Setup snap "docker" (3051) security profiles for auto-connections

......................................................................
Run install hook of "docker" snap if present

2025-01-24T15:55:10Z ERROR run hook "install": cannot perform operation: mount -t tmpfs /tmp/snap.rootfs_eLyHAI: Permission denied

$ sudo snap install docker --channel=latest/edge
error: cannot perform the following tasks:
- Run install hook of "docker" snap if present (run hook "install": cannot perform operation: mount -t tmpfs /tmp/snap.rootfs_llcWsv: Permission denied)

$ cat /etc/os-release 
NAME="Ubuntu Core"
VERSION="22"
ID=ubuntu-core
PRETTY_NAME="Ubuntu Core 22"
VERSION_ID="22"
HOME_URL="https://snapcraft.io/"
BUG_REPORT_URL="https://bugs.launchpad.net/snappy/"

$ snap info docker
...
snap-id: sLCsFAO8PKM5Z0fAKNszUOX0YASjQfeZ
channels:
  latest/stable:    27.2.0   2024-11-22 (2963) 146MB -
  latest/candidate: ↑                                
  latest/beta:      27.4.1   2025-01-06 (2994) 149MB -
  latest/edge:      27.5.0   2025-01-23 (3051) 149MB -
  core18/stable:    20.10.17 2023-03-13 (2746) 146MB -
  core18/candidate: ↑                                
  core18/beta:      ↑                                
  core18/edge:      ↑                                

$ sudo snap install docker --channel=latest/beta
error: cannot perform the following tasks:
- Run install hook of "docker" snap if present (run hook "install": cannot perform operation: mount -t tmpfs /tmp/snap.rootfs_zFkq1W: Permission denied)

farshidtz · 2025-01-24T21:04:12Z

The workflow fails because the distro is empty:
https://github.com/canonical/docker-snap/actions/runs/12914058788/job/36013625887#step:5:83 (2 days ago)
https://github.com/canonical/docker-snap/actions/runs/12957271307/job/36145400313#step:5:83 (now)

This reverts commit cb78626.

farshidtz · 2025-01-25T08:22:20Z

Successful run: https://github.com/canonical/docker-snap/actions/runs/12959400493/job/36160117804

farshidtz

@locnnil please review my changes and merge if you agree. Thanks

Note that PR description is out of date.

locnnil

Looks good! Thank you very much @farshidtz.

locnnil added the CI Continuous integration label Jan 20, 2025

locnnil mentioned this pull request Jan 20, 2025

CI: Run Nvidia workflow on UC22 #216

Closed

locnnil force-pushed the fix-hardcoded-os branch 2 times, most recently from d853657 to 2ac873f Compare January 20, 2025 16:46

locnnil requested review from farshidtz and removed request for farshidtz January 20, 2025 20:50

farshidtz requested changes Jan 22, 2025

View reviewed changes

.github/workflows/nvidia-test.yml Outdated Show resolved Hide resolved

.github/workflows/nvidia-test.yml Outdated Show resolved Hide resolved

.github/workflows/testflinger/nvidia-job.yaml Outdated Show resolved Hide resolved

.github/workflows/nvidia-test.yml Outdated Show resolved Hide resolved

farshidtz mentioned this pull request Jan 22, 2025

Nvidia support core24 #203

Open

locnnil force-pushed the fix-hardcoded-os branch from 2eb834a to fe07e06 Compare January 22, 2025 14:59

locnnil added 5 commits January 22, 2025 14:46

CI: Run Nvidia workflow on UC22

e3db16d

[skip ci] Signed-off-by: Lincoln Wallace <[email protected]>

feat: Loop over OS options and submit TF jobs [skip ci]

5d9b807

Signed-off-by: Lincoln Wallace <[email protected]>

fix: wrong array separator

98f84c7

Signed-off-by: Lincoln Wallace <[email protected]>

refact: change variable name

3d8d4c1

Signed-off-by: Lincoln Wallace <[email protected]>

refact: rename part

3c04e0e

Signed-off-by: Lincoln Wallace <[email protected]>

locnnil force-pushed the fix-hardcoded-os branch from fe07e06 to 3c04e0e Compare January 22, 2025 17:47

farshidtz marked this pull request as draft January 24, 2025 16:55

farshidtz force-pushed the fix-hardcoded-os branch 2 times, most recently from a23ebf0 to f9cdb77 Compare January 24, 2025 17:12

Changes to run tests on UC22

5271976

* Wait for snap changes * Wait for docker daemon after reboot * Run docker commands with sudo * Docker run in non-interactive mode

farshidtz force-pushed the fix-hardcoded-os branch from f9cdb77 to 5271976 Compare January 24, 2025 17:13

Add distro and update example in readme

dd731a5

farshidtz and others added 4 commits January 24, 2025 23:22

Correct indentation for distro field

2517bb0

test: add debug info for distro field [skip ci]

cb78626

Revert "test: add debug info for distro field [skip ci]"

dfb9a13

This reverts commit cb78626.

fix: export DISTRO! [skip ci]

652eb61

farshidtz marked this pull request as ready for review January 25, 2025 08:22

farshidtz approved these changes Jan 25, 2025

View reviewed changes

locnnil commented Jan 25, 2025

View reviewed changes

locnnil merged commit 69554a0 into main Jan 25, 2025
3 of 4 checks passed

locnnil deleted the fix-hardcoded-os branch January 25, 2025 22:23

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

CI: Run Nvidia workflow on UC22 #217

CI: Run Nvidia workflow on UC22 #217

locnnil commented Jan 20, 2025 •

edited

Loading

farshidtz commented Jan 24, 2025

farshidtz commented Jan 24, 2025 •

edited

Loading

farshidtz commented Jan 24, 2025 •

edited

Loading

farshidtz commented Jan 25, 2025

farshidtz left a comment •

edited

Loading

locnnil left a comment

	setup_core() {
	sudo snap install nvidia-core22
	sudo snap install nvidia-assemble --channel 22/stable
	}

	setup() {
	. /etc/os-release

	install_docker

	if [[ $ID == "ubuntu" ]]; then
	setup_classic

	elif [[ $ID == "ubuntu-core" ]]; then
	setup_core

CI: Run Nvidia workflow on UC22 #217

CI: Run Nvidia workflow on UC22 #217

Conversation

locnnil commented Jan 20, 2025 • edited Loading

farshidtz commented Jan 24, 2025

farshidtz commented Jan 24, 2025 • edited Loading

farshidtz commented Jan 24, 2025 • edited Loading

farshidtz commented Jan 25, 2025

farshidtz left a comment • edited Loading

Choose a reason for hiding this comment

locnnil left a comment

Choose a reason for hiding this comment

locnnil commented Jan 20, 2025 •

edited

Loading

farshidtz commented Jan 24, 2025 •

edited

Loading

farshidtz commented Jan 24, 2025 •

edited

Loading

farshidtz left a comment •

edited

Loading