-
-
Notifications
You must be signed in to change notification settings - Fork 1.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
500: Failed to fetch dynamically imported module #3308
Comments
Having the same issue with cloudflare adapter, but just on reload. |
I don't think this happens with all projects or we'd probably be getting a whole lot more bug reports. Can you make it happen with a brand new project? If not, can you provide a project that reproduces it? |
Thanks, @benmccann! Yes, you are probably right - sorry for my assumption. Additionally I have noticed, the first error I receive when loading the page is: But of course I understand that it is hard for you to reproduce it like that. I was just hoping that someone has already experienced the same issue and has a solution. I will keep you updated. Thank you! |
Here is the repo: https://github.com/mawoka-myblock/svelte-cf-error |
Just for a small bit of added context, we are deployed to our own VMs on Digital Ocean (plain old node adapter) and see this error in the logs approximately 1-2 per week. Still trying to gather enough of the situation to share. :( |
@benmccann It also happens on/with the clodflare-pages adapter |
Getting the same error on a hello world app: https://github.com/ManuelDeLeon/sveltekit-firebase-ssr It works as expected locally in dev and preview. The problem is that when deployed to Vercel (https://sveltekit-firebase-ssr.vercel.app/) SSR doesn't work (page source isn't what's expected), an error flashes, and the console shows another error. Error on page:
Error in console
|
Vercel works for me fine, but Cloudflare-Pages doesn't |
Just hit the same error with Cloudflare pages. If you visit the link directly, you get a |
It turns out that in my case, it was because I had to dynamically load certain modules that are only for browsers inside onMount. |
This is actually something we see daily and it's been driving me nuts because I can't seem to find the root cause. In one user session just a bit ago this happened.
Which lead to the 500. I've long thought it was vercel just not rendering those files or an ad tracker but I would love to fix it. Any ideas on how to hunt down the why on this? |
Change to another adapter! The problem may occur only with the vercel adapter, or just with the cloudflare adapter |
Well, I've used both the auto adapter and the vercel one. I'm not picking up and moving off of vercel, unless that's really the issue which means a huge infrastructure change. Now our vercel setup is behind cloudflare and set up the way they recommend. |
You can just deploy it as a test to cloudflare, shouldn't take long and if the issue is still there, it's probably in your code and if not, this issue is the right place! |
I can't even consistently reproduce it, without being able to do that moving it doesn't make any sense. If I could repo or knew how to, I would know where to start looking. On top of that, it's never "just" a just and our site isn't small https://github.com/ThatConference/that.us |
I think that this error might arise whenever you push a new update to the site, which causes a re-built of all modules. In the rebuilt, all files get a new name. If the user was already on the website and doesn't hit rephresh, navigating in the website will eventually try to load a module that is no longer there, causing an error. |
Good point! I think we are having many different issues in this issue, which all lead to the sane error. For me, I fixed my website, by switching to vercel, away from cloudflare. |
Yes, I think that's likely and could cause this error. That should be fixed by #3412. Can folks report whether this issue is fixed by |
That sounds plausible, but could be refuted by our experience. We see this error type ~ one a week and our sessions are incredibly short-lived, certainly much shorter than our deploy cycle. Also, we are using the We will try to report findings once upgraded to |
@benmccann looks like I am on 204 having said that I am working through updating 254. Just trying to make my way through some of the breaking changes. Hoping to get that out maybe this morning. |
@benmccann - I don't have the problem with the latest @sveltejs/kit (1.0.0-next.260) + @sveltejs/adapter-vercel (1.0.0-next.40). Don't get any errors and SSR works like it does locally. |
@benmccann and bumped to 260. Took me a bit longer to work through the breaking changes than I thought. We did see it a few times when testing, but things were in a crazy state of chaos. We will see what happens in the next 24. just for reference for others that might be looking to see what changes we had to make - ThatConference/that.us#1177 |
It didn't fix it for me... |
Unfortunately it happened again last night, and our night traffic is light. Only happened for one user and it appears to have happened a few times for them. We will see what happens today. |
It may still happen for a few days, if users had left their tab open with a previous version of your app. Some users leave tabs open for days or weeks and this can sometimes be the case even for short lived pages such as forms or login pages. A way to verify this sort of bugs in the future, would be to include a variable containing the version of the app in the Sentry report. This will allow to check if the error occurred with a client version that was pushed before the bug fix. |
I'm sorry for the delay here, but I'm happy to report I haven't seen this error surface again. Over the past 72 hours, we've seen 8k page views and 47k requests. @nikosgpet that was an excellent suggestion and one of which I will get added. |
Thanks for the update! I think I'm going to consider this issue closed. If anyone's still running into it, please file an issue with steps to reproduce the error |
Sorry, that was not intentional to spam here. It turns out that it was the usage of {
"adapter-vercel@next": "1.0.0-next.44",
"@sveltejs/kit@next": "1.0.0-next.282"
} After reading their doc, it's not recommended to use in production. I put it here as it might help others. |
Hi got the latest sveltekit, got this error Failed to fetch dynamically imported module: |
Me too. After bump from 175 to 326 I got this error. Page (that is in error path) works only if I remove most of the imports (see the screenshot). Uncommenting any single string throws the error:
|
For me, it was an adblocker blocking some resources (i.e., there was an "ads" somewhere in the URI). |
HOLY MOLY! That was it! |
For us, the issue was that we have the deployment hosted on vercel, and the proxy option turned on in our Cloudflare DNS, and Cloudflare wasn't aware when there was a new deployment. This was causing intermittent issues that resulted in errors described in this post. This was really hard to narrow down, but once we turned off the proxy option in Cloudflare the issues immediately cleared up. This Cloudflare issue is actually described right in the vercel docs here: https://vercel.com/guides/using-cloudflare-with-vercel |
For me, it was "metrics" in the filename. Thank god I figured this out! |
I tracked this down to a Cloudflare Firewall rule, if you set the WAF to "Essentially Off", give it a couple. Test it out it all seems to start working. |
Describe the bug
I am encountering some issues when deploying my project to vercel using the vercel-adapter. The project builds but then the deployed site I receive
500 Failed to fetch dynamically imported module: https://xxx.vercel.app/_app/undefined
The vercel build is triggered by a new Gitlab commit.
Does someone have an idea what could cause this issue? The weird thing is that it works with
npm run build
but not on the deployed vercel site...Reproduction
This should be reproducable by using the same sveltekit + adapter versions
Logs
No response
System Info
Severity
blocking all usage of SvelteKit
Additional Information
No response
The text was updated successfully, but these errors were encountered: