Matt-Yorkley
commented
2 years ago Hi @jsia9 :wave:
Deploying the latest version directly on top of an old release from November 2020 is tricky; there have been too many major changes that have happened in the last 13 months.
You might need to try deploying incremental updates in smaller steps. I made some notes here on milestones, starting from around December 2020:
v3.4.5 (Rails 4.2 and Ruby 2.3.7)
---
v3.5.0 bumps Ruby to 2.4.4
v3.5.11 has a sizeable adjustments migration...
v3.5.12 good_migrations gem was added...
v3.5.14 last major release before the Rails 5.0 upgrade
v3.6.0-rc1 complete Rails 5.0 upgrade
v3.6.1 is the offical Rails 5.0 release. also has adjustments schema changes
v3.6.2 hotfix on v3.6.1 (nils in variant serializer bug?)
---
v3.6.3 bumps Ruby to 2.5.8
v3.6.4 emergency hotfix on v3.6.3 for a customer balance issue
v3.6.5 adds some adjustments schema changes...
v3.6.6 hotfix on v3.6.5... relevant
v3.6.7 some adjustments schema changes...
v3.6.10 last release before Rails 5.2 upgrade
v3.7.0 upgrades to Rails 5.2
v3.7.2 switches to Redis
---
v3.7.3 bumps Ruby to 2.5.9, switches to Sidekiq for jobs, last release before Rails 6.0
v3.8.0 upgrades to Rails 6.0
---
v3.8.3 upgrades to Ruby 2.7.3 and contains the biggest adjustments migrations
v3.8.4 upgrades to Rails 6.1 and adds Webpacker
v3.8.5 switches to Dotenv
v3.8.8 final release before v4.0.0
---
v4.0.0 switches from Unicorn to Puma!
v4.1.0 changes sessions cookie handling...You'll need to provision/deploy with the ofn-v3 branch of ofn-install until you get to v3.8.8. You'll probably need to update the ruby_versions value in inventory/group_vars/all.yml as well when approaching the Ruby version changes...
jsia9
Description
hi, I am trying to upgrade the instance for a demo. I do this to keep the server up to date even though we do not have any active customers. The server was offline the most half of the year and the last upgrade was Nov 2020. So I ran the site.yml and it failed at the deploy.yml with an error that seems common but without any known fix. Maybe because in my case, the instance does not have any orders and therefore the DB instance doest not have any rows to migrate. But here is the error. commenting out the sections as I heard does not work because it will fail at the following step. I commented
- name: clear transient data- name: migrate databaseand it resulted in thisExpected Behavior
Actual Behaviour
This is what happened when running it normally. it seems to be related to the
goodmigrationgem that was added but I am not sure.Steps to Reproduce
Setup an instance in 3.xx version , and run the site.yml.
1. 2. 3. 4.
Severity
bug-s2: a non-critical feature is broken, no workaround
the server has been off for over 2 months while trying to fix the issue. https://github.com/openfoodfoundation/openfoodnetwork/wiki/Bug-severity -->
Your Environment
Ubuntu 18.04.6 LTS (GNU/Linux 5.4.0-1064-azure x86_64) on azure