After upgrading to UMP 8.51, the UMP USM portlet hangs in a loading state for very long periods of time the first time a user logs in. This can take from 20 minutes to an hour for the portlet to load. A query similar to the following is seen running on the MySQL database server:
FROM CM_DEVICE dev
CM_CONFIGURATION_ITEM ci ON ci.dev_id = dev.dev_id
CM_CONFIGURATION_ITEM_METRIC cim ON cim.ci_id = ci.ci_id
CM_CONFIG_ITEM_TO_MASTER meci ON meci.ci_id = ci.ci_id
and (dev.dev_id IN ( 'D043BA3B5CB3B253453AA98BD717FF42A','D5323bd4bb84c2502a8ff28336cdb2710'.,<.....>);
This occurs when there are groups defined in USM with a large number of devices. The list of devices contained in the IN () group in the above query is very large (500+ in this instance).
A non-unique index for the column name ‘ci_id’ for the table ‘CM_CONFIG_ITEM_TO_MASTER’ table was missing after the upgrade.
Add the missing index to the 'CM_CONFIGURE_ITEM_TO_MASTER' table:
create index cm_config_item_to_master_idx01 on cm_config_item_to_master(ci_id);
There is no need to restart the UMP wasp probe after making the change to the table in the MySQL database.