👋 Hello, hello! This is a brand new channel dedicated to discussing and sharing tips about "Prisma-in-production", such as: • Optimizing queries • Making the most out of metrics • Discussing on how migrations are integrated into their CI/CD pipelines, etc. ❓How are you using Prisma in production and what is working / not working? 🤔

Are there any limitations to the prisma data proxy? Are there any compliance concerns for things like soc, hipaa, etc?

Hey @Nurul, we've updated our prisma version and have started recording metrics. When prisma reports losing a connection to the database, we the metrics show we generally have 1 open connection and 17 idle connections. If it was timeout related, would it be a specific timeout error? Any next steps here we can try to get more info?

Hi everyone! I’m running into an issue when using migrations in my CI. The problem is that after the migrations are executed, the Prisma Client seems not to be updating. I’m using GCP, and the deployment is performed at the last stage of my pipeline. However, the Prisma Client seems not to be updated, even though I’m generating it in my build script from

package.json

. In

gitlab-ci.yaml,

the 3 steps are executed in this order: 1. Build (which includes generating the Prisma Client with the latest version of the schema) 2. Migrate 3. Deploy to GCP In `gitlab-ci.yaml`:

build:
  script:
    ...
    - npm install
    - npm run build

...
migrate:
  script:
    - npx prisma migrate deploy

...
deploy:
  script:
    ...
    - gcloud --project "$GCLOUD_PROJECT_ID" app deploy app.yaml

At first, we tried to solve it by adding

npx prisma generate

to the build command from our `package.json`:

"scripts": {
    "build": "npx prisma generate && tsc -p tsconfig.build.json",
...

However, this didn’t work either. Any ideas on what could be happening? GCP seems to cache the node_modules between consecutive deployments if nothing has changed in package.json, but adding the Prisma Client generation to the build script doesn’t solve the issue either… ----------------------------------------------------------- EDIT (2022-10-13): ℹ️ We sorted this out by versioning our

package.json

. Some platforms such as GCP cache the node_modules folder between deployments where

package.json

hasn’t changed at all (not the dependencies, the version or any other change). When you add versioning, your node modules are installed again on your remote GCP instance, and your Prisma Client will be generated.

Hi, I'm having some trouble running migrations from docker. I've got an express app running prisma in one container, and a postgres db in another container. The express app connects fine to the database, but when I try to run

npx prisma migrate deploy

from the inside of that same container I get a connection error. It is very strange because both processes are spawned with the same

DATABASE_URL

environment variable. I've tried running it from a shell from the inside of the container, as well as doing it from a bash script at startup but nothing was worked for me so far. Am I doing something wrong?

specifically, the error I'm getting is:

Error: P1001: Can't reach database server at `sibico-postgres`:`5432`

Please make sure your database server is running at `sibico-postgres`:`5432`.

Yet all my queries in my web app work.

I'm pretty sure this is a problem with prisma because I also have another container running python that too connects correctly to the database to populate it.

It seems like it could be this.

Yeah, that fixed it.

Man, I spent hours debugging this

I am adding prisma to an existing production project (migrating from typeorm). When I followed the instructions here https://www.prisma.io/docs/guides/database/developing-with-prisma-migrate/add-prisma-migrate-to-a-project#baseline-your-production-environment in an attempt to initialize a migration history without applying the migration, it nonetheless deleted all data in my dev database. Am I looking at the wrong documentation? What is the procedure to add prisma without data loss?

Hello @Steven Kuck 👋 Thank you so much for opening the GitHub Issue and providing a reproduction. I can see that the issue was triaged by one of my colleagues, so this is on our engineering team's radar, and our team will be investigating and fixing it in the upcoming Prisma releases.

Hi everyone. Firstly I want to thank you for the great product! I have a task at hand when I have to modify existing Postgres enum. I need to split one enum value into two new values and update existing rows. When I use prisma migrate, it generates a migration where first the new enum type is created, then it is switched in the existing column type, then the old type is removed and the newly created one is renamed to the original name. However this way I can’t modify existing rows because the new value is not valid yet and after the change, the old values are lost. One way would be to first add the two new values, then change existing rows and then create another migration to delete the old enum value. Or I can use two simple SQL statements:

ALTER TYPE "SomeType" RENAME VALUE 'OTHER' TO 'OTHER_PRIVATE';
ALTER TYPE "SomeType" ADD VALUE 'OTHER_TEAM' AFTER 'OTHER_PRIVATE';

Is there some downside to using these

ALTER TYPE

statements over the workflow generated automatically by Prisma? Maybe some compatibility issues? I’m new to Prisma so I’m not sure what’s a better way. Thanks!

I'm not sure if this is the correct channel. Is the prisma team looking at updating the documentation for the Prisma Cloud and Slack Intergration

Hey, I am trying to do a prisma migration deploy on my production database, however it fails reporting an error:

node@worker:/$ node_modules/prisma/build/index.js migrate deploy
Prisma schema loaded from prisma/schema.prisma
Datasource "db": MySQL database "test-database" at "onprem-xtradb-cluster-haproxy.my-organization.svc.cluster.local:3306"

5 migrations found in prisma/migrations

Error: Percona-XtraDB-Cluster prohibits use of GET_LOCK with pxc_strict_mode = ENFORCING
   0: migration_core::state::ApplyMigrations
             at migration-engine/core/src/state.rs:185

database: Percona (mysql) prisma & client: 4.2.1 I understand the error message, but is it possible to bypass it somehow? Quite a few other projects are now running on this database instance and other orm's have no problems with migrations. What am I doing wrong?

Hello 👋 I have this bug in my production build (using docker compose):

[22:05] sapi fast-qr.com-dockercompose $ docker compose build --no-cache
[+] Building 384.5s (14/25)
 => CACHED [internal] load git source <https://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@github.com/erwanvivi>  0.9s
 => [internal] load metadata for <http://docker.io/library/node:alpine|docker.io/library/node:alpine>                                               0.0s
 => CACHED [runner  1/15] FROM <http://docker.io/library/node:alpine|docker.io/library/node:alpine>                                                 0.0s
 => CACHED [runner  2/15] WORKDIR /app                                                                       0.0s
 => [builder 3/6] COPY . .                                                                                   1.2s
 => [runner  3/15] RUN addgroup -g 1001 -S nodejs                                                            1.1s
 => [deps 2/5] RUN apk add --no-cache libc6-compat                                                           1.9s
 => [runner  4/15] RUN adduser -S nextjs -u 1001                                                             1.4s
 => [deps 3/5] WORKDIR /app                                                                                  0.2s
 => [deps 4/5] COPY package.json ./                                                                          0.2s
 => [deps 5/5] RUN npm install --frozen-lockfile                                                           225.9s
 => [runner  5/15] COPY .env* ./                                                                             0.6s
 => [builder 4/6] COPY --from=deps /app/node_modules ./node_modules                                          9.1s
 => ERROR [builder 5/6] RUN npx prisma generate                                                            134.2s
------
 > [builder 5/6] RUN npx prisma generate:
#0 3.869 Prisma schema loaded from prisma/schema.prisma
#0 134.0 Error: request to <https://binaries.prisma.sh/all_commits/461d6a05159055555eb7dfb337c9fb271cbd4d7e/linux-musl/libquery_engine.so.node.sha256> failed, reason: connect ETIMEDOUT 52.222.158.58:443
------
failed to solve: executor failed running [/bin/sh -c npx prisma generate]: exit code: 1

I have found that re-running just after works sometimes just fine:

[22:16] sapi fast-qr.com-dockercompose $ docker compose build --no-cache
[+] Building 768.7s (26/26) FINISHED
 => CACHED [internal] load git source <https://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@github.com/erwanvivi>  0.8s
 => [internal] load metadata for <http://docker.io/library/node:alpine|docker.io/library/node:alpine>                                               0.0s
 => CACHED [deps 1/5] FROM <http://docker.io/library/node:alpine|docker.io/library/node:alpine>                                                     0.0s
 => CACHED [builder 2/6] WORKDIR /app                                                                        0.0s
 => [runner  3/15] RUN addgroup -g 1001 -S nodejs                                                            1.0s
 => [builder 3/6] COPY . .                                                                                   0.9s
 => [deps 2/5] RUN apk add --no-cache libc6-compat                                                           2.1s
 => [runner  4/15] RUN adduser -S nextjs -u 1001                                                             0.8s
 => [runner  5/15] COPY .env* ./                                                                             0.4s
 => [deps 3/5] WORKDIR /app                                                                                  0.2s
 => [deps 4/5] COPY package.json ./                                                                          0.2s
 => [deps 5/5] RUN npm install --frozen-lockfile                                                           212.9s
 => [builder 4/6] COPY --from=deps /app/node_modules ./node_modules                                         11.2s
 => [builder 5/6] RUN npx prisma generate                                                                   10.7s
 => [builder 6/6] RUN npm run build && npm install --omit=dev --ignore-scripts --prefer-offline            480.6s
 => [runner  6/15] COPY --from=builder /app/next.config.js ./                                                0.2s
 => [runner  7/15] COPY --from=builder /app/public ./public                                                  0.7s
 => [runner  8/15] COPY --from=builder --chown=nextjs:nodejs /app/.next ./.next                              1.3s
 => [runner  9/15] COPY --from=builder /app/node_modules ./node_modules                                      9.6s
 => [runner 10/15] COPY --from=builder /app/package.json ./package.json                                      0.1s
 => [runner 11/15] COPY prisma ./prisma                                                                      0.4s
 => [runner 12/15] RUN echo -e '#!/bin/sh\n npx prisma migrate deploy && \\\n node_modules/.bin/next start   0.8s
 => [runner 13/15] RUN chmod +x start.sh                                                                     0.8s
 => [runner 14/15] RUN chown nextjs:nodejs start.sh                                                          0.9s
 => [runner 15/15] RUN chown -R nextjs:nodejs node_modules/@prisma                                           2.1s
 => exporting to image                                                                                      12.5s
 => => exporting layers                                                                                     12.5s
 => => writing image sha256:2e5e83514dbfa0894cad69ada1ddbd6bf87f0f19957a28211b48a25570e0205b                 0.0s
 => => naming to <http://docker.io/library/fast_qrcom-fastqr|docker.io/library/fast_qrcom-fastqr>                                                         0.0s

I have found this issue : https://prisma.slack.com/archives/CA491RJH0/p1653083645405259 Which is kinda not resolved, can you help me ?