Merge pull request #564 from jburke-cadc/CADC-13234

CADC-13234 fix ringhold for ArtifactDAO API change

.github/workflows/gradle.yml

@@ -70,7 +70,6 @@ jobs:
     - name: java build -- ratik
       run: cd ratik && ../gradlew --info clean build javadoc checkstyleMain
 
-## disabled until updated for cadc-inventory-db API changes
-#    - name: java build -- ringhold
-#      run: cd ringhold && ../gradlew --info clean build javadoc checkstyleMain
+    - name: java build -- ringhold
+      run: cd ringhold && ../gradlew --info clean build javadoc checkstyleMain
 

ringhold/README.md

@@ -1,7 +1,8 @@
-# Storage Inventory local artifact removal process (ringhold)
+# Storage Inventory local artifact deletion process (ringhold)
 
-Process to remove local artifacts that are no longer being synchronised by fenwick. This tool is used
-to perform quick cleanup at a storage site after changing the fenwick artifact-filter policy.
+Process to remove the local copy of artifacts from a storage site inventory database and 
+generate DeletedStorageLocationEvent(s) so the removal will propagate correctly to a global inventory. 
+This does not remove the files from storage (see `tantar`).
 
 ## configuration
 See the [cadc-java](https://github.com/opencadc/docker-base/tree/master/cadc-java) image docs for general config requirements.
@@ -18,17 +19,25 @@ org.opencadc.ringhold.inventory.schema={schema for inventory database objects}
 org.opencadc.ringhold.inventory.username={username for inventory admin}
 org.opencadc.ringhold.inventory.password={password for inventory admin}
 org.opencadc.ringhold.inventory.url=jdbc:postgresql://{server}/{database}
+
+# artifact namespace(s) to remove
+org.opencadc.ringhold.namespace={storage site namespace}
+
+# artifact uri bucket filter (optional)
+org.opencadc.ringhold.buckets={uriBucket prefix or range of prefixes}
 ```
 The `inventory` account owns and manages all the content (insert, update, delete) in the inventory schema. Unlike
 other components that modify inventory content, this component **does not initialise** the database objects because
 it never makes sense to run this in a new/empty database. The database is specified in the JDBC URL. Failure to 
 connect to a pre-initialised database will show up in logs.
 
-### artifact-deselector.sql
-Contains a SQL clause used as a WHERE constraint. The clause returns Artifact's that match the URI pattern.
-```
-WHERE uri LIKE 'cadc:CFHT/%'
-```
+The `namespace` is the prefix of the Artifact URI's to be deleted. The `namespace` must end with a colon (:) 
+or slash (/) so one namespace cannot accidentally match (be a prefix of) another namespace. Multiple values 
+of `namespace` may be specified, one per line.
+
+The `buckets` value indicates a subset of artifacts to delete. The range of uri bucket prefixes is specified 
+with two values separated by a single - (dash) character; whitespace is ignored. Multiple instances of `ringhold` 
+can be run (in parallel) to subdivide the work as long as the range of buckets do not overlap.
 
 ## building it
 ```

ringhold/VERSION

@@ -1,6 +1,6 @@
 ## deployable containers have a semantic and build tag
 # semantic version tag: major.minor[.patch]
 # build version tag: timestamp
-VER=0.2.0
+VER=0.3.0
 TAGS="${VER} ${VER}-$(date --utc +"%Y%m%dT%H%M%S")"
 unset VER