SYS.REMOVE_METASTORE_ENTRY¶
This procedure drops a table from the external catalog if it exists (without checking that the table exists in the catalog). However, it does not handle related policies and base tables; hence they must be dropped separately.
You must connect to the server from the Snappy shell, check the policies and base tables, and drop them separately.
Syntax¶
Example¶
Considering the case when cluster fails to come up, and the log mentions the pattern: "AnalysisException: Table (dbName).(tableName) might be inconsistent in hive catalog. Use system procedure SYS.REMOVE_METASTORE_ENTRY to remove inconsistency", then use the following steps to resolve the issue.
-
Check if the server is running:
cd $SNAPPY_HOME ./sbin/snappy-status-all.sh
-
Launch Snappy shell and connect to the server:
./bin/snappy ; connect client '<server hostname>:<server port>';
-
Check and find if there are any policies on the table which caused catalog inconsistency and drop policies:
SELECT * FROM SYS.SYSPOLICIES; DROP POLICY <policy name>;
-
Call the procedure to drop the table from the catalog:
call sys.REMOVE_METASTORE_ENTRY('<dbName>.<tableName>', 'false');
-
Restart SnappyData cluster and check the status. The cluster starts successfully.
Also see: