Closed tomdavies closed 4 years ago
This migration comes from craft core
@ostark , sorry maybe I wasn’t clear. I know it’s a core migration, but the migration breaks if this plugin is installed/enabled. If the plugin is disabled the migration works, so it must be something in the plugin code that causes the error, no?
Sorry, I can not imagine a relation
Hello,
I'd like to say I'm experiencing almost exactly the same problem that @tomdavies is describing. I am upgrading Craft from 3.3.6 => 3.4.12 and craft-async-queue from 1.4 => 2.1.
I am getting an almost identical stack trace as @tomdavies as it looks like some async-queue code is running (See stack trace on 4 and 5) during the Craft migration.
If I disable the async-queue plugin before running the Craft migration scripts, the migration completes successfully. I can then re-enable the craft-async-queue plugin on the upgraded site.
Ah, it's a chicken egg problem. The migration adds a new 'channel' field to the queue table and fires an event that injects a job to the queue which makes use of the new field.
https://github.com/ostark/craft-async-queue/releases/tag/2.1.1 thanks, this should fix it
Thanks for the quick fix @ostark! I've run my migration scripts locally a few times now and that update seems to be fixing it for me.
Cheers! 🍻
Yeah I had a similar thing with Neo and the same upgrade. Events are a pain the arse
Tom Davies tom@tomdavies.net
On 3 Apr 2020, at 20:05, Oliver Stark notifications@github.com wrote:
Ah, it's a chicken egg problem. The migration adds a new field and fires a event that injects a job to the queue which makes use of the new field.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub, or unsubscribe.
Hey @ostark 👋,
Plugin version: 2.1.0
Not sure if this a bug, but thought I'd report it just in case. I hit an error when attempting to upgrate a site that uses async-queue from Craft 3.3.20.1 => 3.4.11.
When running
./craft migrate/all
I got this error (full stack trace after):Disabling async-queue and running again (albeit with a fresh database dump) worked, so it's not blocking me any more, just thought you should know about it.
Happy to try and make a reduced test case if that's helpful....
Stack trace: