Compose deploy fails: relation "patch" does not exist

[bnqtoan]

Description

When deploying a compose project via API or webhook, the deployment fails with a database error.

Error

Failed query: select "patchId", "type", "filePath", "enabled", "content", "createdAt", "updatedAt", "applicationId", "composeId" from "patch" "patch" where "patch"."composeId" = $1 order by "patch"."filePath" asc
params: jNp_z5Yt5ugtj0Jjq9k_v
Error occurred ❌, check the logs for details.

Steps to Reproduce

1. Create a compose project with GitHub source
2. Trigger deploy via API (POST /api/compose.deploy) or webhook
3. Deploy fails after cloning repo successfully

Expected Behavior

Deploy should complete successfully after cloning.

Environment

• Dokploy version: Latest (installed via official script ~March 4, 2026)
• Docker Swarm mode
• Compose project with GitHub source type

Logs

The error appears in Dokploy container logs:

Cloning Repo github.com/xxx/xxx.git to /etc/dokploy/compose/xxx/code: ✅
Failed query: select ... from "patch" where "patch"."composeId" = $1
Error occurred ❌, check the logs for details.

Workaround

Running docker compose up -d manually in the code directory works fine.

[dosubot]

Hi @bnqtoan! I'm Dosu and I’m helping the dokploy team.

This error occurs because the patch table is missing from your database. This table was added in v0.28.0 (released Feb 27, 2026) via migration 0145, which should have run during installation but likely failed silently.

Why this happens: Dokploy's migration script catches errors but doesn't halt startup, so the server runs with an incomplete schema. Docker Swarm installations are particularly prone to this due to networking issues where the install script misconfigures the advertise address.

To diagnose, check if migrations failed:

docker service logs dokploy

Look for errors like Migration failed Error: connect ECONNREFUSED.

Recommended fix (if Swarm networking is the issue):

docker swarm leave --force
docker swarm init --advertise-addr <YOUR_PRIVATE_IP>
curl -sSL https://dokploy.com/install.sh | sh

Quick workaround - manually create the table:

docker exec -it $(docker ps -q -f name=dokploy-postgres) psql -U dokploy -d dokploy

Then run:

CREATE TYPE "public"."patchType" AS ENUM('create', 'update', 'delete');
CREATE TABLE "patch" (
    "patchId" text PRIMARY KEY NOT NULL,
    "type" "patchType" DEFAULT 'update' NOT NULL,
    "filePath" text NOT NULL,
    "enabled" boolean DEFAULT true NOT NULL,
    "content" text NOT NULL,
    "createdAt" text NOT NULL,
    "updatedAt" text,
    "applicationId" text,
    "composeId" text,
    CONSTRAINT "patch_filepath_application_unique" UNIQUE("filePath","applicationId"),
    CONSTRAINT "patch_filepath_compose_unique" UNIQUE("filePath","composeId")
);
ALTER TABLE "patch" ADD CONSTRAINT "patch_applicationId_application_applicationId_fk" FOREIGN KEY ("applicationId") REFERENCES "public"."application"("applicationId") ON DELETE cascade ON UPDATE no action;
ALTER TABLE "patch" ADD CONSTRAINT "patch_composeId_compose_composeId_fk" FOREIGN KEY ("composeId") REFERENCES "public"."compose"("composeId") ON DELETE cascade ON UPDATE no action;

After either fix, restart Dokploy: docker service scale dokploy=0 && docker service scale dokploy=1

To reply, just mention @dosu.

---

Share context across your team and agents. Try Dosu.

[Siumauricio]

@bnqtoan by some reason the migrations didn't execute automatically

You can restart the service running this commands

docker service scale dokploy=0
docker service scale dokploy=1

[Siumauricio]

Closed due to inactivity