Really appreciate the help, trying to
Really appreciate the help, trying to come up with a solution so that I can leverage Convex for its realtime capabilities and use Postgres for the relational side of things
27 Replies
The error indicates that the Prisma Client was generated for a different runtime environment than the one required during deployment. Specifically, the client was generated for "debian-openssl-1.1.x" but the deployment environment requires "debian-openssl-3.0.x".
To resolve this, update the binaryTargets in the schema.prisma file to include "debian-openssl-3.0.x" and regenerate the Prisma Client by running prisma generate.
This ensures the correct query engine is available at runtime. Additionally, verify that the Prisma Client is being generated in the expected output path and that no bundler or deployment configuration (such as optimizePackageImports in Next.js) is interfering with the inclusion of the query engine file
^ that's an AI generated result from google
Hmm, curious
Let's give it a shot
Same result unfortunately
I'm thinking that it might be the convex environment
Yeah me too
That's making it behave wierd
Ensure that you ran
prisma generate and that "libquery_engine-debian-openssl-3.0.x.so.node" has been copied to "data/tmp/.tmphP04YY/source/841c2d5a-2c11-4a5a-bfd9-2a2621f8d6c3/modules/_deps/node".
This is a bit of a clue😅
There's prisma accelerate but isn't that another cost?
Don't know much about that, sorry!
Maybe just leave it in next api route then?
The only solution i can think of is to create a separate environment for it, and access it with Http
I've got one sentence: wtf
I'm wondering if something like airbyte would help, or to break out of prisma and sql and move all to convex depending on how large your project is
I started off with using only Convex
But things quickly slowed down as I tried to import 12,000+ rows of data and perform JOIN queries
And then there's the damn "too many bytes read in a single function" warning
Which is inherently due to the fact that convex is again not optimized for OLAP
I see i see
Are you using a monorepo?
Yeah
Nice and organized
If you've got an api folder in your app use that to access the HTTP requests, if not, create a separate api package with prisma and hono, and create your requests there, and place the hono api url to the .site of convex, if that didn't work, use it under your localhost:3001/api or something like that
Hono is very small and fast enough to not cost you delays
The current approach right now is Next API routes that are able to work with Prisma properly
That's the only thing i can think of, or airbyte
Nice! Great to hear
I think ultimately you're right though that API calls will be the "escape hatch" for convex in this edge case
Normally that's actions in "use node"
But even that can't handle whatever it is that prisma is doing to generate their client
https://clickhouse.com/clickhouse plus this thing
ClickHouse
Real-Time Data Analytics Platform | ClickHouse
Get real-time insights with ClickHouse, the high-performance data analytics platform. Reap the benefits of streamlined data analysis and try for free today.
Tbf prisma is scuffed
Yeah that's why I was so thrilled about convex
I haven't worked with it for 2 years 😂 i always go with drizzle if i had to
I never got the chance to dive into drizzle
Ideally I don't want to have to bother with any of that
That's also a good point
Convex Ents might be another approach, but it's not officially backed by convex so there are some concerns with stability and continued support