syndesis: [Upgrade] After performing a rollback, upgraded (and rolled back) deployments don't start

This is a…


[ ] Feature request
[ ] Regression (a behavior that used to work and stopped working in a new release)
[X] Bug report  
[ ] Documentation issue or request

The problem

I’m performing the rollback in upgrade_50_replace_template, where I’m appending exit 1 after update_version $tag in replace_template::run().

After everything is rolled back, the deployments are scaled back to 1, but only those, that are not affected by upgrade (prometheus, oauthproxy, db [i’m not doing db migration for now]) are started back.

This is how it looks like after rollback: uz3h

The completed deployments in my case can’t be scaled up - when I scale them via oc or via web, it goes back to 0 after few seconds.

cc @rhuss @paoloantinori @dsimansk

Expected behavior

Screenshot

Request and Response Data

API Endpoints and Schemas

Tasks involved / Steps to Reproduce

About this issue

Original URL
State: closed
Created 6 years ago
Comments: 16 (8 by maintainers)

Commits related to this issue

fix(upgrade): Fix rollback fixes #2499 — committed to rhuss/syndesis by rhuss 6 years ago

Most upvoted comments

Running integrations should not be affected by an infrastructure upgrade. Since the sampledb can indeed be used by those, I wonder whether we should exclude backup/restore completely for this DB ? I mean, at the end its a “sample” DB anyway. @paoloantinori @zregvart any opinions on this ?

It’s a sample database we could probably remove it from next release and create a sample database template, like we do for AMQ, so not backing it up is perfectly OK with me.

zregvart on May 2, 2018