deployment issues
Hi! I’ve deployed my project on fly.io but the issue is beyond the landing page auth isn’t working. I followed the doc instructions. Perhaps the Prisma DB didn’t deploy? Is there an easy way to see the log and debug?
My deployment: https://supastage-client.fly.dev
Thanks for the help.
AI Enhanced Interior Design
Interior Designers using AI to create stunning virtual staging.
44 Replies
Hi @Arnab. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:
----------------------
Sure, I can guide you on how to debug this issue.
Firstly, you can check the network tab in your browser's developer tools to see if requests to the server are taking a long time or failing.
Secondly, you can check the logs on your Fly.io server. To do this, you can use the
flyctl logs command. This will show you what is being printed on the server, whether it's getting requests from the client, and if it's reporting any errors.
If you suspect that the issue might be with the Prisma DB, you can connect to your Fly production database and interact with it using Prisma Studio. To do this, you can open up a tunnel with fly proxy 5432 -a <db-name>. After setting up the tunnel, you can execute wasp db studio to interact with the database. Just remember to remove the DATABASE_URL with the production database URL from .env.server if you put it there, so you don't accidentally do something you didn't want to on the production database later.
Here are the relevant sources for your reference:
- Discord support thread: My login tab on the frontpage went missing after deployment on fly.io > Answers Part 2 (https://discord.com/channels/686873244791210014/1217941277979574315)
- Discord support thread: How to interact with database in prisma studio after deployment? > Answers (https://discord.com/channels/686873244791210014/1220423321422663820)
- Docs > Advanced > Deployment > Deploying Manually > Fly.io (server and database) > Deploy to a Fly.io App (https://wasp-lang.dev/docs/advanced/deployment/manually#deploy-to-a-flyio-app)
I hope this helps! Let me know if you have any other questions.
----------------------
Note: you can ask me a follow up question by @ mentioning me again :speech_balloon:
----------------------
Sorry @Arnab, I seem to be having a bit of trouble. Don't worry though, I've pinged my support and they will try to fix me up! :technologist:
Perhaps try again in 15mins?So I just gave it a quick try, and I see that the server is not responding, requests are taking super long time and never happen.
Was the app working at any point, or is this from the start since you deployed?
Can you check the logs on Fly for the server? What about logs for the database?
One common problem is that people forget to provide all the env vars for the server on Fly, so the server doesn't start properly.
Thanks for the help. It's working locally. But it's never worked after deploying.
The logs don't have errors. The env vars are likely the issue. I put the info in the .env.server in the project.
Is there a step that I'm missing or how do I make sure that the env vars are properly moved over?
Let's figure this out, couple of questions:
1. Could you still copy the logs that you have, so we can see what they look like, for the server on Fly? There might be something there that is not obvious at the first glance.
2. How did you deploy to Fly, did you use these instructions https://wasp-lang.dev/docs/advanced/deployment/cli ?
3. Server vars -> so it will not pick up vars from .env.server, those are only for your local development. You need to set env vars on the Fly server yourself, those are vars you will use for production. I realize now this is not mentioned explicitly in the instrutions above hm, I wonder if we should add that to that piece of docs. Here is a piece about it from the "manual deployment" section of the docs: https://wasp-lang.dev/docs/advanced/deployment/manually#environment-variables .
Here's the logs:
Yes I followed the instructions on the CLI. I
3. Ahh ok that makes more sense. I understood that wasp deploy command took care of all that.
Also do you guys recommend hosting the DB on fly.io or some other service? I'm trying to keep it simple, cheap and quick to deploy
If you want to go for simple and quick, I would say stick to Fly at the moment, due to Wasp's
wasp deploy fly functionality. It is also not expensive, although not free unfortunately.
We will be looking to provide more options in the future that are this streamlined and hopefully some that are free, but for now this sounds like the easiest one.
And yeah, also host DB there then.
As for your issue -> did setting up the env vars help?
As for logs from the server -> ok, so these logs are not that useful, true, because the useful ones are probably quite old. What you can try is redeploy your app on Fly (using wasp deploy fly deploy), and then, right after deployment was successful, go checkout the logs of the server again. You should see more useful logs, something like "server is starting", and potentially any interesting errors also.Trying it now. Solid help! Sorry I'm a bit of a noob.
I'm still getting errors after updating the envs:"
2024-05-22T14:36:31Z proxy[e286de4f920486] mia [error][PM05] failed to connect to machine: gave up after 15 attempts (in 16.413149496s)
2024-05-22T14:36:31Z proxy[e286de4f920486] mia [error][PM05] failed to connect to machine: gave up after 15 attempts (in 16.458412933s)
"
Also what configuration do I need for the host machines, and for which (server, db, client)? Right now I'm on the development shared.
Ok so I redeployed the project. I also set the secrets as you suggested. But login isn't working! arrghhh
Here's the logs from server, client, and db right after deploying and trying to login.
I'd appreciate the help
From the logs it seems like server keeps restarting.... which is weird. Is it because I'm on a 1 shared?
Also another noob question. All the .env.server keys etc are they all automatically added or do I need to add them somehow?
Hey no worries, the whole deployment thing is realtively complicated, even for me with quite some experience!
OK, so indeed your Fly machine for the supastage-server is complaining a lot about restarting machine and stuff, I am nto sure what are all those errors about (that is on the Fly side). You could check with their support for help on that. Maybe take a look at metrics (cpu, memory), is that looking all good.
As for the logs, I found the error in the supastage-server logs:
From this, I am guessing your Sendgrid API key is not set to a proper value, but to a mock value, and that is why server won't start.
So .env.server remains on your machine, doesn't goa nywhere. It is only for development.
On Fly, you need to set the env vars manually. We set some of them for you if you are using
wasp deploy fly deploy, specifically DATABASE_URL, and JWT_SECRET, and PORT, but the rest you need to set on your own.
So for example, if you need Sendtring key, you will need to set that manually on Fly, specifically for your supastage-server machine.
You can do this like this:
and do it multiple times for any secret you need to set. You can also check which secrets you already have set, by running wasp deploy fly cmd --context server secrets list
Which docs were you following for all this so far? I guess https://wasp-lang.dev/docs/advanced/deployment/cli ? I am looking at them again and they are really lacking this info on setting server env vars, that sucks, I will make sure we improve that!
Ok, I created an issue for it here, we will be improving these docs soon! https://github.com/wasp-lang/wasp/issues/2053Amazing help! I really appreciate this.
I'm trying it now.
I was using the Opensaas docs, which have even less lol.
Wohooo @Arnab, you just became a Waspeteer level 3!
Hm you are right, docs in open-saas give just enough info to make you feel like you know all that is needed to deploy, but not enough to actually deploy successfully!
I also opened an issue for that, we will make sure to fix this! https://github.com/wasp-lang/open-saas/issues/139
Thanks a lot for the effort and patience here, this will help others 🙂
Thank you! OK I pushed all the envs to the server and I'm still getting server errors.
Also, when I try to do the google auth (https://supastage-server.fly.dev/auth/google/login) I get the page isn't working
More logs "2024-05-23T17:20:03.893 proxy[148e057c001189] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-05-23T17:20:12.473 proxy[0806d4ef615058] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-05-23T17:20:14.479 proxy[148e057c001189] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-05-23T17:20:22.489 proxy[0806d4ef615058] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-05-23T17:20:24.495 proxy[148e057c001189] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-05-23T17:20:31.503 proxy[148e057c001189] lga [error] [PR04] could not find a good candidate within 90 attempts at load balancing"
If you happen to be online now. I could jump on a google meet, instead of the back and forth. There's got to be something stupid I'm messing up lol
did you add the URL of your Server to your
Authorized Redirct URI's section in Google Cloud? https://wasp-lang.dev/docs/auth/social-auth/google#3-creating-a-google-oauth-appGoogle | Wasp
Wasp supports Google Authentication out of the box.
Did you aslo set the
JWT_SECRET environment variable on the server?Yes I set JWT. Yes on google auth also
Hey guys I appreciate your help here. Any chance we could jump on a quick call to fix the issue? It may be faster, and I'm trying to deploy asap.
Yes
@martinsos any chance you can help with the deployment issues?
@Arnab which version of wasp are you on?
if you're on 0.12 or lower it should be
.../auth/login/google , but if you're on 0.13 it should be .../auth/google/callback
also could you share your server logs when you login with google. Here's what a successful login looks like in v0.13
0.13.2
I've already shared the logs above. Nothing about google auth is showing up.
I keep getting the same errors
"2024-06-04T15:38:39.219 proxy[0806d4ef615058] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-06-04T15:38:47.140 proxy[148e057c001189] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-06-04T15:38:47.156 proxy[148e057c001189] dfw [error] [PR04] could not find a good candidate within 90 attempts at load balancing
2024-06-04T15:38:50.181 proxy[0806d4ef615058] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-06-04T15:38:57.393 proxy[148e057c001189] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-06-04T15:39:01.020 proxy[0806d4ef615058] ewr [error] [PC01] instance refused connection. is your app listening on 0.0.0.0:8080? make sure it is not only listening on 127.0.0.1 (hint: look at your startup logs, servers often print the address they are listening on)
2024-06-04T15:39:04.309 proxy[148e057c001189] lga [error] [PR04] could not find a good candidate within 90 attempts at load balancing
Am I making a mistake with the type of instance?
I'm using "shared-cpu-1x@1024MB"
@martinsos I see that you may be online to help?
I've been trying to fix this issue for a few weeks. I'd really apprecaite the help.
Also in the oAuth do I keep it on testing or prod? If it's on prod doesn't it need a custom domain?
Let's jump on a call tomorrow and try to figure it out. I'm free 9 am to 2 pm CEST
thank you!
I'm in EDT. I think 2pm CEST is 8am EST. I can do 1:30pm CEST if needed.
Awesome! THanks @miho . Sorry for not responding sooner @Arnab , I was swamped, but also, @miho will be able to offer much better help, he knows this stuff quite better than me!
Let's try 1:30 if possible 👍 send me a DM and I'll send you a meet link
Awesome! Talk soon
Thanks for the help
@miho still free in 10 min?
Yep, let's do it
Hi @Arnab, Any updates? I hope you figured it out. I'm having the same issue 🙂
Yes, @miho helped fix the issue.
The issue I ran into was with one of the secrets. The EMAIL_PASSWORD secret I had set via terminal and forgot the "". I'd go to fly.io and set it directly on their dashboard.
Also you can add a "console.log (emailUsername, emailPassword)" after line 35 for logs in actions.ts.
Wohooo @Arnab, you just became a Waspeteer level 4!
Also make sure your port secret is 8080
Oh, great to hear, congrats! Thank you for your feedback, I'll try it, and also good luck with your product launch!
We also updated the
PORT env var on Fly to be 8080 😊Thank you @miho , I'll give it a try, and provide feedback 😃
It worked, thank you guys (@Arnab @miho @martinsos ) ❤️
Nice! What was your exact issue?
I created a special issue just for this, because our system for reporting missing env vars in production is not great and having a better thign in place could help quite a bit: https://github.com/wasp-lang/wasp/issues/2086 .
I still didn't quite get what the issue was though: @Arnab , you had EMAIL_PASSWORD secret set in .env.server, for local development, but you forgot to set it on Fly.io, is that it?
What do you mean with adding a console.log?
Port -> @miho don't we set that for them when deploying to Fly? You somehow changed its value by accident @Arnab so that caused additional issues?
@rdm what was your problem?
Hey @martinsos I think the issue is that I'm a noob coder lol. The issue that miho helped with is that I pushed all the envs via terminal. But when you are sending a password with spaces (Google email password) it screws it up if quotes aren't included. So I set it directly on the fly dashboard.
It was a very simple issue, that I think a more experience coder would know. Maybe as a suggestion make it easier to set secrets/ pass along envs?
I also suggested to miho to record a video of deploying on youtube. I think videos on youtube would be super helpful in general. I'm happy to do a quick video and share. Lmk
@Arnab I could imagine myself easily making the same mistake! At Wasp we really don't want people to have to bother with such details, so I would certainly like us to help with this somehow.
Those are good ideas! I would formalize it like this:
1. Allow people to define production env variables via .env.production if they wish. Although this needs to be done properly on our side, so it is a bit more work for us, especially desiging the experinec here to be right.
2. Improve the docs, by also warning people about quotes and giving more clear examples on how to do it.
3. Finally add video as you said! Videos are on one hand very valuable, but on the other hand tricky to maintain because if sometihng cahnges, editing / redoing video is much more expensive then editing text, so that is why we don't have a lot of them. Well, initial production is also more time consuming. But having one video of deployment process, that might be reasonable. If you would like to do a quick video that would be cool, absolutely!
Ok, here is the issue where we are tracking this issue, I will make sure we prioritize it in the following sprints!
We could also add instructions to our deployment docs for adding the env vars on Fly.io's dashboard -- this is how I also did it for Open SaaS 🙂