An error occurred collecting the necessary information to run your Deploystack
dvdvdvdvdvd opened this issue · 20 comments
The project is not found? What permissions are denied?
Help Token: AfeSHlKhbIC_i-0rG2wIGB2jDkHOQDjvnWYoxZ_zLbVxfJbc08tSZAX9MblMSBFlLKPA7vMcGxOTHxpgWPvCen-PHS_9jz_jgnEGRdEg2aswu0dt
I have the same issue as the other reported.
Appreciate the help!
I tried deploying in a brand new GCP account and I got this error, a bit different.
Thanks for reporting this. I can confirm the problem is the project name has spaces. (This is a bug, deploystack should be able to handle that).
I'm working on a fix.
In the meantime, you can workaround by creating a project through the deploystack interface by choosing "CREATE NEW PROJECT" option 1.
OR
Creating a project in the Cloud Console with no spaces in the name.
Sorry about that.
I tried those two things, and they also failed, even creating a new project. In the first image there are no spaces in the "sandbox" project. Thank you kindly!
Welp, now I got two problems. ;)
I'll investigate more. Thanks for your patience.
Oh quick question. Is this your own individual account, or are you using this on your employer provided GCP account? (Wondering if there are org restrictions that might be in play.)
In the first image, it's my employers account.
In the second image, it's my own account, brand new, full access. I thought that might be the issue too, but it seems to be something else.
Sorry for the delay. There was a delay in getting the latest deploystack cloud shell image into the default cache...
The behavior of Deploystack should change to wait for human prompting before making calls that need authorization. This MIGHT make things better. I'm not entirely certain. There are few other changes that I'm working on.
Bad news, good news... then unfortunately, some more bad news. But we're getting close!
Bad news: it didn't work when I tried creating a new project.
Good news: I used the project "testingoncemore" which I manually enabled billing for since I wanted to trying a few other things using Cloud Run, and that worked! (I also noticed "threetiertest" which was the name of the new project I tried to create above was listed as "(Billing Disabled)")
Bad news: Even though I waited a few minutes the todo app seemed to have issues caching data to Redis and it timeout. I tried refreshing several times but no luck.
(This is my personal GCP account just for testing so I don't mind sharing URLs or IP addresses.)
Sounds right.
- The Redis server looks good.
- The secrets look okay too.
Where are my manners... I really appreciate your help Terry! Your video and documentation were exactly what I was looking for last week. What you put together is awesome, my team and I want to use this architecture to upgrade our current web app so I'm super excited about it. (no pressure ;) )
Let me know if there's anything else I can do to help.
Okay, so I'm doing some digging.
The health check that should work on the api layer isn't working:
https://todo-api-hjqhtfgepq-uc.a.run.app/healthz
I think that means that it isn't even init'ing correctly.
It looks like Redis is setup right and the secret is properly picked up by Cloud Run, so it's likely a networking issue.
And I just manually reran in and I am in a broken state. Which means my tests are passing even when stuff is broken - not cool. So.. bear with me.
And thank you so much for the kind words! I'm glad this would be helpful for you - if the engineer wasn't so bug producing. :)
Okay. This particular stack is now working. I added tests that would catch this scenario - all the resources spun up, but the app didn't work.
Now on to the billing stuff.
This has been fixed in the core, it will take about a week to get into the Open In Cloud Shell base image.
Sounds great! I'm going to try to deploy this again tomorrow, hopefully it's all fine.
Everything deployed but the app doesn't work. You said it could take up to a week so I'll try again on Thursday.
(Apply complete! Resources: 38 added, 0 changed, 0 destroyed.
Outputs:
api = "https://three-tier-app-api-t5bzevkwsq-uc.a.run.app"
endpoint = "https://three-tier-app-fe-t5bzevkwsq-uc.a.run.app"
sqlservername = "three-tier-app-db-3be7")
Awesome, appreciate all your help!
I noticed that too, (The Same Origin Policy disallows reading the remote resource at https://api/v1/todo.)
Separately,
- https://three-tier-app-api-t5bzevkwsq-uc.a.run.app/api/v1/todo does work.
- https://three-tier-app-api-t5bzevkwsq-uc.a.run.app/healthz doesn't work.
Idk if this is to be expected.
Getting closer and closer!
It works! https://three-tier-app-fe-oxk25f6pnq-uc.a.run.app/
- I tried to creating a new project and that failed.
I enabled billing manually and tried again with:
deploystack install - This time there was an error starting cloud run api and it said to try again
I didn't make any modification and just ran
deploystack installon the same project
Then it worked.
I'm going to clone the repo and start to see how I can use this as a template for our new project. I want to switch out Golang for Python, MySQL for Postgres and I want to use Nuxt.js on the frontend. I know in your video you said this is taken care of in the CI/CD layer just by changing the docker files. If you have any other helpful tips that'd be great!
Thanks for your patience!