Add harvest & xloader worker configuration

- Update Dockerfiles - Update start_ckan.sh with custom workers - Update start_ckan_development.sh.override to remove automatic workers - Add supervisor harvester.conf - Add supervisor xloader.conf - Update Scheming DCAT version to 2.1.0
mjanez · Mar 8, 2024 · 68a9ce9 · 68a9ce9
1 parent 0ed33f2
commit 68a9ce9
Show file tree

Hide file tree

Showing 8 changed files with 83 additions and 32 deletions.
diff --git a/.env.example b/.env.example
@@ -89,6 +89,7 @@ CKAN_SITE_URL=http://localhost:81
 CKAN__ROOT_PATH=/catalog/{{LANG}}
 CKAN_PORT=5000
 CKAN__FAVICON=/catalog/base/images/ckan.ico
+CKAN__SITE_LOGO=/images/default/ckan-logo.png
 CKAN___BEAKER__SESSION__SECRET=CHANGE_ME
 # See https://docs.ckan.org/en/latest/maintaining/configuration.html#api-token-settings
 CKAN___API_TOKEN__JWT__ENCODE__SECRET=string:CHANGE_ME
@@ -125,7 +126,7 @@ CKAN__LOCALE_ORDER="en es pt_BR ja it cs_CZ ca fr el sv sr sr@latin no sk fi ru
 CKAN__LOCALES_OFFERED="en es pt_BR ja it cs_CZ ca fr el sv sr sr@latin no sk fi ru de pl nl bg ko_KR hu sa sl lv"
 
 # Extensions
-CKAN__PLUGINS="envvars stats text_view image_view webpage_view recline_view resourcedictionary datastore xloader harvest ckan_harvester spatial_metadata spatial_query spatial_harvest_metadata_api csw_harvester waf_harvester doc_harvester resource_proxy geo_view geojson_view wmts_view shp_view dcat dcat_rdf_harvester dcat_json_harvester dcat_json_interface scheming_dcat_datasets scheming_dcat_groups scheming_dcat_organizations scheming_dcat scheming_dcat_ckan_harvester scheming_dcat_csw_harvester pdf_view pages fluent"
+CKAN__PLUGINS="envvars stats text_view image_view webpage_view recline_view resourcedictionary datastore xloader harvest spatial_metadata spatial_query spatial_harvest_metadata_api csw_harvester waf_harvester doc_harvester resource_proxy geo_view geojson_view wmts_view shp_view dcat dcat_rdf_harvester dcat_json_harvester dcat_json_interface scheming_dcat_datasets scheming_dcat_groups scheming_dcat_organizations scheming_dcat scheming_dcat_ckan_harvester scheming_dcat_xls_harvester pdf_view pages fluent"
 
 # ckanext-harvest
 CKAN__HARVEST__MQ__TYPE=redis
@@ -137,7 +138,7 @@ CKAN__HARVEST__LOG_TIMEFRAME=40
 
 # ckanext-xloader
 CKANEXT__XLOADER__API_TOKEN=api_token
-CKANEXT__XLOADER__JOBS__DB_URI=postgresql://ckan:ckan@db/ckan
+CKANEXT__XLOADER__JOBS__DB_URI=postgresql://ckandbuser:ckandbpassword@db/ckandb
 
 # ckanext-dcat
 CKANEXT__DCAT__BASE_URI=${CKAN_URL}

diff --git a/ckan/Dockerfile b/ckan/Dockerfile
@@ -22,7 +22,7 @@ COPY req_fixes req_fixes
 ## Pages - v0.5.2 ##
 ## PDFView - 0.0.8 ##
 ## Fluent - v1.0.1  (Forked stable version) ##
-## Scheming DCAT - v2.0.0 (GeoDCAT-AP/NTI-RISP extended version) ##
+## Scheming DCAT - v2.1.0 (GeoDCAT-AP/NTI-RISP extended version) ##
 RUN echo ${TZ} > /etc/timezone && \
     if ! [ /usr/share/zoneinfo/${TZ} -ef /etc/localtime ]; then cp /usr/share/zoneinfo/${TZ} /etc/localtime; fi && \
     # Remove apk cache
@@ -54,8 +54,8 @@ RUN echo ${TZ} > /etc/timezone && \
     echo "mjanez/ckanext-fluent" && \
     pip3 install --no-cache-dir -e git+https://github.com/mjanez/ckanext-fluent.git@v1.0.1#egg=ckanext-fluent && \
     echo "mjanez/ckanext-scheming_dcat" && \
-    pip3 install --no-cache-dir -e git+https://github.com/mjanez/ckanext-scheming_dcat.git@v2.0.0#egg=ckanext_scheming_dcat && \
-    pip3 install --no-cache-dir -r https://raw.githubusercontent.com/mjanez/ckanext-scheming_dcat/v2.0.0/requirements.txt
+    pip3 install --no-cache-dir -e git+https://github.com/mjanez/ckanext-scheming_dcat.git@v2.1.0#egg=ckanext_scheming_dcat && \
+    pip3 install --no-cache-dir -r https://raw.githubusercontent.com/mjanez/ckanext-scheming_dcat/v2.1.0/requirements.txt
 
 # Used to configure the container environment by setting environment variables, creating users, running initialization scripts, .etc
 COPY docker-entrypoint.d/* /docker-entrypoint.d/
@@ -79,8 +79,8 @@ RUN for d in $APP_DIR/patches/*; do \
 COPY setup/start_ckan.sh.override ${APP_DIR}/start_ckan.sh
 RUN chmod +x ${APP_DIR}/start_ckan.sh
 
-## Harvester
-COPY setup/workers/harvester.conf /etc/supervisord.d/harvester.conf
+## Load workers supervisor configuration
+COPY setup/workers/* /etc/supervisord.d/
 
 # Start CKAN
 CMD ["/bin/sh", "-c", "$APP_DIR/start_ckan.sh"]
diff --git a/ckan/Dockerfile.dev b/ckan/Dockerfile.dev
@@ -60,8 +60,8 @@ COPY setup/who.ini ./
 COPY setup/start_ckan_development.sh.override ./start_ckan_development.sh
 RUN chmod +x ./start_ckan_development.sh
 
-## Harvester
-COPY setup/workers/harvester.conf /etc/supervisord.d/harvester.conf
+## Load workers supervisor configuration
+COPY setup/workers/* /etc/supervisord.d/
 
 # Apply any patches needed to CKAN core or any of the built extensions (not the
 # runtime mounted ones)

diff --git a/ckan/Dockerfile.ghcr b/ckan/Dockerfile.ghcr
@@ -38,4 +38,13 @@ RUN for d in $APP_DIR/patches/*; do \
     fi ; \
 done
 
+# Workers
+## Update start_ckan.sh with custom workers
+COPY setup/start_ckan.sh.override ${APP_DIR}/start_ckan.sh
+RUN chmod +x ${APP_DIR}/start_ckan.sh
+
+## Load workers supervisor configuration
+COPY setup/workers/* /etc/supervisord.d/
+
+# Start CKAN
 CMD ["/bin/sh", "-c", "$APP_DIR/start_ckan.sh"]
diff --git a/ckan/setup/start_ckan.sh.override b/ckan/setup/start_ckan.sh.override
@@ -6,7 +6,6 @@ if grep -E "beaker.session.secret ?= ?$" ckan.ini
 then
     echo "Setting beaker.session.secret in ini file"
     ckan config-tool $CKAN_INI "beaker.session.secret=$(python3 -c 'import secrets; print(secrets.token_urlsafe())')"
-    ckan config-tool $CKAN_INI "WTF_CSRF_SECRET_KEY=$(python3 -c 'import secrets; print(secrets.token_urlsafe())')"
     JWT_SECRET=$(python3 -c 'import secrets; print("string:" + secrets.token_urlsafe())')
     ckan config-tool $CKAN_INI "api_token.jwt.encode.secret=${JWT_SECRET}"
     ckan config-tool $CKAN_INI "api_token.jwt.decode.secret=${JWT_SECRET}"
@@ -51,17 +50,18 @@ UWSGI_OPTS="--plugins http,python \
 if [ $? -eq 0 ]
 then
     # Start supervisord
+    echo "[prerun.workers] Loading the CKAN workers with supervisord..."
     supervisord --configuration /etc/supervisord.conf &
+
+    # Workers
+    ## Add harvester background procces to crontab
+    echo "[prerun.workers] Add harvester background procceses to crontab"
+    crontab -l | { cat; echo "*/15 * * * * /usr/bin/supervisorctl start ckan_harvester_run"; } | crontab -
+    ## Clean-up mechanism for the harvest log table. 'ckan.harvest.log_timeframe'. The default time frame is 30 days
+    crontab -l | { cat; echo "0 5 */30 * * /usr/bin/supervisorctl start ckan_harvester_clean_log"; } | crontab -
+
     # Start uwsgi
     uwsgi $UWSGI_OPTS
 else
   echo "[prerun] failed...not starting CKAN."
 fi
-# Workers
-## Start the Harvester worker
-echo "[prerun.workers] Starting the CKAN Harvester worker"
-ckan harvester run
-## Add harvester to crontab
-crontab -l | { cat; echo "0 */2 * * * ckan harvester run"; } | crontab -
-## Clean-up mechanism for the harvest log table. 'ckan.harvest.log_timeframe'. The default time frame is 30 days
-crontab -l | { cat; echo "0 5 * * * ckan harvester clean-harvest-log"; } | crontab -
diff --git a/ckan/setup/start_ckan_development.sh.override b/ckan/setup/start_ckan_development.sh.override
@@ -51,7 +51,6 @@ if grep -E "beaker.session.secret ?= ?$" ckan.ini
 then
     echo "Setting beaker.session.secret in ini file"
     ckan config-tool $CKAN_INI "beaker.session.secret=$(python3 -c 'import secrets; print(secrets.token_urlsafe())')"
-    ckan config-tool $CKAN_INI "WTF_CSRF_SECRET_KEY=$(python3 -c 'import secrets; print(secrets.token_urlsafe())')"
     JWT_SECRET=$(python3 -c 'import secrets; print("string:" + secrets.token_urlsafe())')
     ckan config-tool $CKAN_INI "api_token.jwt.encode.secret=${JWT_SECRET}"
     ckan config-tool $CKAN_INI "api_token.jwt.decode.secret=${JWT_SECRET}"
@@ -95,15 +94,13 @@ mkdir -p $CKAN_LOGS_PATH/xloader
 chown -R ckan:ckan $CKAN_LOGS_PATH/xloader
 
 # Start supervisord
-supervisord --configuration /etc/supervisord.conf
-
-# Workers
-## Start the Harvester worker
-echo "[prerun.workers] Starting the CKAN Harvester worker"
-ckan harvester run
-
-## Clean-up mechanism for the harvest log table
-ckan harvester clean-harvest-log
+#supervisord --configuration /etc/supervisord.conf &
 
 # Start the development server as the ckan user with automatic reload
-su ckan -c "/usr/bin/ckan -c $CKAN_INI run -H 0.0.0.0"
+su ckan -c "/usr/bin/ckan -c $CKAN_INI run -H 0.0.0.0"
+
+# Workers
+# To start the Harvester worker
+# ckan harvester run
+# Clean-up mechanism for the harvest log table
+# ckan harvester clean-harvest-log
diff --git a/ckan/setup/workers/harvester.conf b/ckan/setup/workers/harvester.conf
@@ -3,17 +3,49 @@ command=ckan harvester gather-consumer
 user=ckan
 numprocs=1
 stdout_logfile=/var/log/harvester/gather_consumer.log
-stderr_logfile=/var/log/harvester/gather_consumer.err.log
+stdout_logfile_maxbytes=50MB
+stderr_logfile=/var/log/harvester/gather_consumer.log
+stderr_logfile_maxbytes=50MB
 autostart=true
 autorestart=true
 startsecs=10
+priority=1
 
 [program:ckan_fetch_consumer]
 command=ckan harvester fetch-consumer
 user=ckan
 numprocs=1
 stdout_logfile=/var/log/harvester/fetch_consumer.log
-stderr_logfile=/var/log/harvester/fetch_consumer.err.log
+stdout_logfile_maxbytes=50MB
+stderr_logfile=/var/log/harvester/fetch_consumer.log
+stderr_logfile_maxbytes=50MB
 autostart=true
 autorestart=true
-startsecs=10
+startsecs=10
+priority=2
+
+[program:ckan_harvester_run]
+command=ckan harvester run
+user=ckan
+numprocs=1
+stdout_logfile=/var/log/harvester/ckan_harvester.log
+stdout_logfile_maxbytes=25MB
+stderr_logfile=/var/log/harvester/ckan_harvester.log
+stderr_logfile_maxbytes=25MB
+autostart=true
+autorestart=false
+startsecs=2
+priority=3
+
+[program:ckan_harvester_clean_log]
+command=ckan harvester clean-harvest-log
+user=ckan
+numprocs=1
+stdout_logfile=/var/log/harvester/ckan_harvester_clean_log.log
+stdout_logfile_maxbytes=25MB
+stderr_logfile=/var/log/harvester/ckan_harvester_clean_log.log
+stderr_logfile_maxbytes=25MB
+autostart=false
+autorestart=false
+startsecs=2
+priority=4
diff --git a/ckan/setup/workers/xloader.conf b/ckan/setup/workers/xloader.conf
@@ -0,0 +1,12 @@
+[program:ckan_xloader]
+command=ckan jobs worker default
+user=ckan
+numprocs=1
+stdout_logfile=/var/log/harvester/ckan_xloader.log
+stdout_logfile_maxbytes=100MB
+stderr_logfile=/var/log/harvester/ckan_xloader.log
+stderr_logfile_maxbytes=100MB
+autostart=true
+autorestart=true
+startsecs=4
+priority=1