backfilling taking a very very long time.

dusty-caterpillar-550 is taking a very very long time doing this:

⠧ Backfilling indexes and checking that documents match your schema...

⠧ Backfilling indexes and checking that documents match your schema...

13 Replies

ian•2y ago

Did it resolved by waiting or did you restart the process? Chatted offline. It took a minute to deploy but eventually succeeded. Sorry it was slow for you

magicsethOP•2y ago

🙂 It's taken over 2 minutes right now and still no sign of completion backfilling I'm starting it over again now it gave me an error I think it might be related to an experimental thing I was trying with ian Hard to understand what is happening actually

✖ Error: Unable to push deployment config to https://dusty-caterpillar-550.convex.cloud
400 Bad Request: InvalidModules: Loading the pushed modules encountered the following
    error:
Not implemented: constructor with body type other than string | ArrayBuffer | Blob | FormData | null for Request. Consider calling an action defined in Node.js instead (https://docs.convex.dev/functions/actions).
  at throwUncatchableDeveloperError (../../udf-system/src/helpers.ts:50:8)
  at throwNotImplementedMethodError (../../udf-system/src/helpers.ts:10:46)
  at new Request (../../udf-system/src/23_request.ts:125:10)
  at <anonymous> (../../node_modules/ky/source/core/constants.ts:11:13)
  at <anonymous> (../../node_modules/ky/source/core/constants.ts:24:7)

✖ Error: Unable to push deployment config to https://dusty-caterpillar-550.convex.cloud
400 Bad Request: InvalidModules: Loading the pushed modules encountered the following
    error:
Not implemented: constructor with body type other than string | ArrayBuffer | Blob | FormData | null for Request. Consider calling an action defined in Node.js instead (https://docs.convex.dev/functions/actions).
  at throwUncatchableDeveloperError (../../udf-system/src/helpers.ts:50:8)
  at throwNotImplementedMethodError (../../udf-system/src/helpers.ts:10:46)
  at new Request (../../udf-system/src/23_request.ts:125:10)
  at <anonymous> (../../node_modules/ky/source/core/constants.ts:11:13)
  at <anonymous> (../../node_modules/ky/source/core/constants.ts:24:7)

It seems that pinecone-client depends on ky but I haven't had any problems deploying in the past and reverting doesn't fix it I think it broke when I first pinged you at 3:38 @ballingt I think something broke in my production deployment also? I'm seeing those errors in productionnow I made a change yesterday that could have caused it, but I don't know how to revert it now Possibly: I have a helpers file that has functions that include pinecone-client Before, I could call those functions from a query because they weren't actually using pinecone-client (other helper functions were) now it seems I can't?

ballingt•2y ago

Oh that sounds like it could be the issue If before it was just actions that imported it but now queries do, that could be a problem

magicsethOP•2y ago

I don't think my code changed Is it possible deployment got stricter?

ballingt•2y ago

@nipunn and I just checked, there have been no changes today well small changes this morning

ian•2y ago

I found that I couldn't use pinecone-client in v8, but "@pinecone-database/pinecone": "^0.1.6", works. And iirc that's what the repo of yours I saw before used? In case that helps. ky depends on setTimeout which we're hoping to support soon, but hasn't worked yet afaik

magicsethOP•2y ago

So it seems like my production deployment broke sometime after 1:21 PM PST today. I was running npx convex dev, and things were working, and then suddenly they were unable to be deployed, and rolling back to previous checkpoints wouldn't work either.

jamwt•2y ago

broke meaning production isn't working, or new deployments aren't pushing?

magicsethOP•2y ago

it was both I couldn't push and the thing that had been working started erroring about the fact that ky isn't supported Sorry I didn't grab the production error

nipunn•2y ago

Oh yeah - @ballingt and I chatted with @magicseth earlier! It seemed like there was definitely some kind of regression where something that previously was working no longer worked. We were able to find the stack trace of the error, so we should be able to investigate and make sure this doesn't happen again.

magicsethOP•2y ago

thanks @nipunn and thanks @ballingt for working on the debugging!

sshader•2y ago

I believe the InvalidModules error should be fixed for your deployments now, but let us know if that's not the case! Sorry for the regression

magicsethOP•2y ago

Thanks @sshader I made my code correct too 🙂 well, correcter

backfilling taking a very very long time.

Did you find this page helpful?