fix(superset): restore ClickHouse dialect and dataset metadata sync

- switch Superset ClickHouse URI back to clickhousedb://\n- refresh dataset metadata during init to restore filter columns\n- install runtime deps in image layer and add troubleshooting notes
This commit is contained in:
2026-02-11 09:23:48 +03:00
parent 9b08b924ef
commit 7819af8688
3 changed files with 27 additions and 6 deletions
+2 -2
View File
@@ -6,6 +6,6 @@ RUN apt-get update && \
apt-get install -y bind9-host bat less iputils-ping curl postgresql-client && \
apt-get clean
USER superset
RUN pip install --no-cache-dir clickhouse-connect==0.8.* psycopg2-binary==2.9.*
RUN pip install clickhouse-connect==0.8.* clickhouse-sqlalchemy==0.3.* psycopg2-binary==2.9.*
USER superset
+8
View File
@@ -293,5 +293,13 @@ curl -s -X POST -u admin:admin http://localhost:3000/api/admin/provisioning/aler
docker volume rm <project>_grafana_lib
docker compose up -d grafana
```
- В Superset ошибки `DB engine Error` и `Cannot load filter`, а в логах есть `Can't load plugin: sqlalchemy.dialects:clickhouse.connect`:
- Причина: некорректный URI диалекта ClickHouse (`clickhouse+connect://...`).
- Используйте URI `clickhousedb://...` и пересоберите сервисы Superset:
```bash
docker compose build superset superset-init
docker compose up -d clickhouse
docker compose up -d --force-recreate superset-init superset
```
- После `docker compose down -v` нужно повторно прогнать: `ddl_init` -> `kafka_load` -> `etl_pipeline`.
- Для демо по умолчанию использовать малый срез данных; полный прогон делать осознанно.
+17 -4
View File
@@ -54,7 +54,7 @@ def build_clickhouse_uri() -> str:
host = CLICKHOUSE_HOST
port = CLICKHOUSE_PORT
database = quote_plus(CLICKHOUSE_DATABASE)
return f"clickhouse+connect://{user}:{password}@{host}:{port}/{database}"
return f"clickhousedb://{user}:{password}@{host}:{port}/{database}"
def run_superset_cli(args):
@@ -139,7 +139,7 @@ def import_datasets():
# Создаем Python скрипт для выполнения внутри superset shell
script_lines = [
"import clickhouse_connect # Регистрирует диалект clickhouse+connect",
"import clickhouse_connect # Регистрирует диалект clickhousedb",
"from superset.extensions import db",
"from superset.models.core import Database",
"from superset.connectors.sqla.models import SqlaTable",
@@ -170,7 +170,18 @@ def import_datasets():
").first()"
),
"if existing:",
f" print('Dataset {table_name} already exists')",
" try:",
" if not existing.columns:",
" existing.fetch_metadata()",
" db.session.commit()",
f" print('Refreshed dataset metadata: {table_name}')",
" refreshed += 1",
" else:",
f" print('Dataset {table_name} already exists')",
" except Exception as e:",
f" print(f'ERROR: failed to refresh {table_name}: {{e}}')",
" db.session.rollback()",
" errors += 1",
"else:",
" try:",
(
@@ -183,8 +194,10 @@ def import_datasets():
")"
),
" db.session.add(dataset)",
" db.session.flush()",
" dataset.fetch_metadata()",
" db.session.commit()",
f" print('Created dataset: {table_name} (metadata will be fetched on first use)')",
f" print('Created dataset: {table_name}')",
" imported += 1",
" except Exception as e:",
f" print(f'ERROR: failed to create {table_name}: {{e}}')",