Storage test suite flakiness

[uncleDecart]

This is a tracker to collect all Storage test suite runs, which failed in GitHub Actions

How to contribute?

If you see any failure in GitHub actions runs, please add it to this table.

How to find eden version?

Open workflow file of the run and see which eden tag is invoked, like here it is 0.9.3.

How to find EVE version?

In setup-job of workflow which failed find eve image parameter which in inputs. For example, here it is evebuild/pr:3698. In case if it contains evebuild/pr put PR in it.

Link to run	Eden version	EVE version	test	file system
1	0.9.5	PR	volume_test	zfs
2	0.9.5	master	volumes_test	zfs
3	master	11.7.0	volumes_test	zfs

How to run suite locally

Clean previous build and add default config with debug level of printing

make clean && make build-tests
./eden config add default
./eden config set default --key eve.log-level --value debug

(optional) set EVE tag to a version you're testing

./eden config set default --key eve.tag --value 11.3.0

Setup and run a test

./eden setup
./dist/bin/eden+ports.sh 2223:2223
./eden start
./eden eve onboard
./eden test ./tests/workflow -s storage.tests.txt -v debug

[eriknordmark]

A recent run https://github.com/lf-edge/eve/actions/runs/7806282543/job/21310964084?pr=3742 shows the ZFS storage test suite failing to deploy an app instance due to insufficient memory. Can we easily increase the memory for the virtual device?

[uncleDecart]

A recent run https://github.com/lf-edge/eve/actions/runs/7806282543/job/21310964084?pr=3742 shows the ZFS storage test suite failing to deploy an app instance due to insufficient memory. Can we easily increase the memory for the virtual device?

I think yes, maybe @giggsoff knows where exactly. I can take a look later this week

[giggsoff]

A recent run https://github.com/lf-edge/eve/actions/runs/7806282543/job/21310964084?pr=3742 shows the ZFS storage test suite failing to deploy an app instance due to insufficient memory. Can we easily increase the memory for the virtual device?

I think yes, maybe @giggsoff knows where exactly. I can take a look later this week

We can add here something like ./eden config set default --key=eve.ram --value=6144. Or we can increase it by default here.

[uncleDecart]

A recent run https://github.com/lf-edge/eve/actions/runs/7806282543/job/21310964084?pr=3742 shows the ZFS storage test suite failing to deploy an app instance due to insufficient memory. Can we easily increase the memory for the virtual device?

I think yes, maybe @giggsoff knows where exactly. I can take a look later this week

We can add here something like ./eden config set default --key=eve.ram --value=6144. Or we can increase it by default here.

In BuildJet we have 16 GBs, but will it work on local forks?.. I'm not so sure. But since we are targeting BuildJet runners, we should be fine

[giggsoff]

A recent run https://github.com/lf-edge/eve/actions/runs/7806282543/job/21310964084?pr=3742 shows the ZFS storage test suite failing to deploy an app instance due to insufficient memory. Can we easily increase the memory for the virtual device?

I think yes, maybe @giggsoff knows where exactly. I can take a look later this week

We can add here something like ./eden config set default --key=eve.ram --value=6144. Or we can increase it by default here.

In BuildJet we have 16 GBs, but will it work on local forks?.. I'm not so sure. But since we are targeting BuildJet runners, we should be fine

The same memory for GH hosted.

[uncleDecart]

Okay so now we fail on other thing but that's progress

[uncleDecart]

Stabilised. Will open new one if something occurs again