You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
When running docker stack deploy to update existing (old) simcore services, the docker services did not fetch the latest docker images for the specified tag. This should be considered a bug and needs a fix. It was mitigated by updating the services with re-pulling docker images via portainer in this case.
When running docker stack deploy, the webserver did at least once revert to an old state, where the docker service of the simcore-webserver did not have a necessary env-var specified for itself. It was in effect missing one env-var. Another run of docker stack deploy has fixed the problem. We are not sure about the root cause. One assumption is that the stack staging-simcore previously existed and was created through the portainer-api. Portainer then assumes that it has "full ownership" over the stack. "Updating" the stack using docker stack deploy might lead portainer to become confused about the desired state of the service. I suggest to next time before releasing running docker stack rm staging-simcore to eleminate this possibility and have the docker stack deploy command recreate the stack from scratch.
The registry (ops-service) on aws-staging had a wrong env-var for the S3 bucket name specified. I did not manage to find out how this wrong variable comes into play. I have manually fixed the variable in portainer. Likely upon the next release to staging this might happen again, beware.
Further notes:
I think we agreed we would not announce staging pre-releases anymore in mattermost etc., since all users for their regular work have been moved to production. So maybe these sections in the pre-release github issue template should be removed
What kind of pre-release?
master branch
Sprint Name
TheNameless
Pre-release version
4
Commit SHA
a03ed85924a9246c62926d6da362d4aee48c2197
Did the commit CI suceeded?
Motivation
Weekly staging release
What Changed
director-v2
env var definition in compose spec by making it explicit #4786 by @GitHKDevops check⚠️ devops
e2e testing check 🧪
No response
Summary 📝
make release-staging name=TheNameless version=4 git_sha=a03ed85924a9246c62926d6da362d4aee48c2197
https://github.com/ITISFoundation/osparc-simcore/releases/new?prerelease=1&target=<commit_sha>&tag=staging_<sprint_name><version>&title=Staging%20<sprint_name><version>
maintenance
in every concerned deployment){"start": "2023-10-03T11:00:00.000Z", "end": "2023-10-03T13:00:00.000Z", "reason": "Release TheNameless4"}
🔊 Maintenance scheduled for NAMED_DAY DD. MM from START_TIME - END_TIME.
@ALL Be aware that you will automatically be logged out and your projects stopped and saved during the maintenance time. Affected:
and on premises:
Reason: Scheduled staging-release of STAGING_NAME_AND_VERSION.
Thanks for your understanding and sorry for the inconveniences,
Your friendly oSparc Team
The text was updated successfully, but these errors were encountered: