Merge pull request #1068 from CREDO23/fix/zero-cache

[Fix] zero-cache startup crash on high-core machines & restrict replication to needed tables

docker/.env.example

@@ -79,6 +79,23 @@ EMBEDDING_MODEL=sentence-transformers/all-MiniLM-L6-v2
 # Change ZERO_ADMIN_PASSWORD for security in production.
 
 # ZERO_ADMIN_PASSWORD=surfsense-zero-admin
+
+# Publication restricting which tables zero-cache replicates from Postgres.
+# Created automatically by Alembic migration 116.
+# Only change this if you manage publications manually.
+# ZERO_APP_PUBLICATIONS=zero_publication
+
+# Sync worker tuning — zero-cache defaults ZERO_NUM_SYNC_WORKERS to the number
+# of CPU cores, which can exceed the connection pool limits on high-core machines.
+# Each sync worker needs at least 1 connection from both the UPSTREAM and CVR
+# pools, so these constraints must hold:
+#   ZERO_UPSTREAM_MAX_CONNS >= ZERO_NUM_SYNC_WORKERS
+#   ZERO_CVR_MAX_CONNS      >= ZERO_NUM_SYNC_WORKERS
+# Default of 4 workers is sufficient for self-hosted / personal use.
+# ZERO_NUM_SYNC_WORKERS=4
+# ZERO_UPSTREAM_MAX_CONNS=20
+# ZERO_CVR_MAX_CONNS=30
+
 # Full override for the Zero → Postgres connection URLs.
 # Leave commented out to use the Docker-managed `db` container (default).
 # ZERO_UPSTREAM_DB=postgresql://surfsense:surfsense@db:5432/surfsense

docker/docker-compose.dev.yml

@@ -176,7 +176,7 @@ services:
     extra_hosts:
       - "host.docker.internal:host-gateway"
     depends_on:
-      db:
+      backend:
         condition: service_healthy
     environment:
       - ZERO_UPSTREAM_DB=${ZERO_UPSTREAM_DB:-postgresql://${DB_USER:-postgres}:${DB_PASSWORD:-postgres}@${DB_HOST:-db}:${DB_PORT:-5432}/${DB_NAME:-surfsense}?sslmode=${DB_SSLMODE:-disable}}
@@ -184,6 +184,10 @@ services:
       - ZERO_CHANGE_DB=${ZERO_CHANGE_DB:-postgresql://${DB_USER:-postgres}:${DB_PASSWORD:-postgres}@${DB_HOST:-db}:${DB_PORT:-5432}/${DB_NAME:-surfsense}?sslmode=${DB_SSLMODE:-disable}}
       - ZERO_REPLICA_FILE=/data/zero.db
       - ZERO_ADMIN_PASSWORD=${ZERO_ADMIN_PASSWORD:-surfsense-zero-admin}
+      - ZERO_APP_PUBLICATIONS=${ZERO_APP_PUBLICATIONS:-zero_publication}
+      - ZERO_NUM_SYNC_WORKERS=${ZERO_NUM_SYNC_WORKERS:-4}
+      - ZERO_UPSTREAM_MAX_CONNS=${ZERO_UPSTREAM_MAX_CONNS:-20}
+      - ZERO_CVR_MAX_CONNS=${ZERO_CVR_MAX_CONNS:-30}
       - ZERO_QUERY_URL=${ZERO_QUERY_URL:-http://frontend:3000/api/zero/query}
       - ZERO_MUTATE_URL=${ZERO_MUTATE_URL:-http://frontend:3000/api/zero/mutate}
     volumes:

docker/docker-compose.yml

@@ -170,13 +170,17 @@ services:
       ZERO_CHANGE_DB: ${ZERO_CHANGE_DB:-postgresql://${DB_USER:-surfsense}:${DB_PASSWORD:-surfsense}@${DB_HOST:-db}:${DB_PORT:-5432}/${DB_NAME:-surfsense}?sslmode=${DB_SSLMODE:-disable}}
       ZERO_REPLICA_FILE: /data/zero.db
       ZERO_ADMIN_PASSWORD: ${ZERO_ADMIN_PASSWORD:-surfsense-zero-admin}
+      ZERO_APP_PUBLICATIONS: ${ZERO_APP_PUBLICATIONS:-zero_publication}
+      ZERO_NUM_SYNC_WORKERS: ${ZERO_NUM_SYNC_WORKERS:-4}
+      ZERO_UPSTREAM_MAX_CONNS: ${ZERO_UPSTREAM_MAX_CONNS:-20}
+      ZERO_CVR_MAX_CONNS: ${ZERO_CVR_MAX_CONNS:-30}
       ZERO_QUERY_URL: ${ZERO_QUERY_URL:-http://frontend:3000/api/zero/query}
       ZERO_MUTATE_URL: ${ZERO_MUTATE_URL:-http://frontend:3000/api/zero/mutate}
     volumes:
       - zero_cache_data:/data
     restart: unless-stopped
     depends_on:
-      db:
+      backend:
         condition: service_healthy
     healthcheck:
       test: ["CMD", "curl", "-f", "http://localhost:4848/keepalive"]

surfsense_backend/alembic/versions/116_create_zero_publication.py

@@ -0,0 +1,52 @@
+"""create zero_publication for zero-cache replication
+
+Restricts zero-cache replication to only the tables the frontend
+queries via Zero, instead of replicating all tables in public schema.
+
+See: https://zero.rocicorp.dev/docs/zero-cache-config#app-publications
+
+Revision ID: 116
+Revises: 115
+"""
+
+from collections.abc import Sequence
+
+import sqlalchemy as sa
+
+from alembic import op
+
+revision: str = "116"
+down_revision: str | None = "115"
+branch_labels: str | Sequence[str] | None = None
+depends_on: str | Sequence[str] | None = None
+
+PUBLICATION_NAME = "zero_publication"
+
+TABLES = [
+    "notifications",
+    "documents",
+    "folders",
+    "search_source_connectors",
+    "new_chat_messages",
+    "chat_comments",
+    "chat_session_state",
+]
+
+
+def upgrade() -> None:
+    conn = op.get_bind()
+    exists = conn.execute(
+        sa.text("SELECT 1 FROM pg_publication WHERE pubname = :name"),
+        {"name": PUBLICATION_NAME},
+    ).fetchone()
+    if not exists:
+        table_list = ", ".join(TABLES)
+        conn.execute(
+            sa.text(
+                f"CREATE PUBLICATION {PUBLICATION_NAME} FOR TABLE {table_list}"
+            )
+        )
+
+
+def downgrade() -> None:
+    op.execute(f"DROP PUBLICATION IF EXISTS {PUBLICATION_NAME}")

surfsense_web/content/docs/docker-installation/docker-compose.mdx

@@ -71,6 +71,10 @@ Defaults work out of the box. Change `ZERO_ADMIN_PASSWORD` for security in produ
 | `ZERO_UPSTREAM_DB` | PostgreSQL connection URL for replication (must be a direct connection, not via pgbouncer) | *(built from DB_* vars)* |
 | `ZERO_CVR_DB` | PostgreSQL connection URL for client view records | *(built from DB_* vars)* |
 | `ZERO_CHANGE_DB` | PostgreSQL connection URL for replication log entries | *(built from DB_* vars)* |
+| `ZERO_APP_PUBLICATIONS` | PostgreSQL publication restricting which tables are replicated (created by migration 116) | `zero_publication` |
+| `ZERO_NUM_SYNC_WORKERS` | Number of view-sync worker processes. Must be ≤ connection pool sizes | `4` |
+| `ZERO_UPSTREAM_MAX_CONNS` | Max connections to upstream PostgreSQL for mutations | `20` |
+| `ZERO_CVR_MAX_CONNS` | Max connections to the CVR database | `30` |
 
 ### Database
 

surfsense_web/content/docs/how-to/zero-sync.mdx

@@ -36,6 +36,10 @@ zero-cache is included in the Docker Compose setup. The key environment variable
 | `ZERO_ADMIN_PASSWORD` | Password for the zero-cache admin UI and `/statz` endpoint | `surfsense-zero-admin` |
 | `ZERO_UPSTREAM_DB` | PostgreSQL connection URL for replication | Built from `DB_*` vars |
 | `NEXT_PUBLIC_ZERO_CACHE_URL` | URL the frontend uses to connect to zero-cache | `http://localhost:<ZERO_CACHE_PORT>` |
+| `ZERO_APP_PUBLICATIONS` | PostgreSQL publication restricting which tables are replicated | `zero_publication` |
+| `ZERO_NUM_SYNC_WORKERS` | Number of view-sync worker processes. Must be ≤ `ZERO_UPSTREAM_MAX_CONNS` and ≤ `ZERO_CVR_MAX_CONNS` | `4` |
+| `ZERO_UPSTREAM_MAX_CONNS` | Max connections to upstream PostgreSQL for mutations | `20` |
+| `ZERO_CVR_MAX_CONNS` | Max connections to the CVR database | `30` |
 
 ### Manual / Local Development
 
@@ -83,6 +87,7 @@ Zero syncs the following tables for real-time features:
 ## Troubleshooting
 
 - **zero-cache not starting**: Check `docker compose logs zero-cache`. Ensure PostgreSQL has `wal_level=logical` (configured in `postgresql.conf`).
+- **"Insufficient upstream connections" error**: zero-cache defaults `ZERO_NUM_SYNC_WORKERS` to the number of CPU cores, which can exceed connection pool limits on high-core machines. Lower `ZERO_NUM_SYNC_WORKERS` or raise `ZERO_UPSTREAM_MAX_CONNS` / `ZERO_CVR_MAX_CONNS` in your `.env`.
 - **Frontend not syncing**: Open DevTools → Console and check for WebSocket connection errors. Verify `NEXT_PUBLIC_ZERO_CACHE_URL` matches the running zero-cache address.
 - **Stale data after restart**: zero-cache rebuilds its SQLite replica from PostgreSQL on startup. This may take a moment for large databases.