Complete the following steps to resolve the issue:

1. Back up the IMA data store.

2. Using DSView, examine the Data Store Properties of the custom evaluator and the contained rules. Load Evaluators could be found at the following branch:
   \ServerNeighborhoods\%FarmName%\LMS\Load Evaluators 
   
   In this example, there are three evaluators displayed. Two of the evaluators begin with 01ff ; these are the built-in Default and Advanced. The custom evaluator node that begins with 4c33 is expanded and its name and the rules assigned.

3. Notice that there are duplicate rules assigned to this evaluator. For the Load Management System (LMS) to work correctly, all assigned rules must be unique.

4. To rectify this, remove the load evaluator and all assigned servers, (known as LEUsers) from the data store using TSQL. Note of the following:

   

• The Context Identifier (CID) of the node. In this case, it is 0x1e6.
  

• The Unique Identifier (UID) of the corrupted evaluator. (4c33-0005-0000d6d9)
  
  The name of the corrupted evaluator.

1. Using SQL Management Studio, connect a query session to the SQL server and database hosting the data store.

2. Obtain the NodeID that correlates to the preceding records. Convert the collected CID values from Hex to Decimal.
   cid 0x1e6 = Context ID 486
   cid 0x1e8 = Context ID 488

3. Execute the following SQL statement and examine the results.
   Note: Repeat this process for each CID.
   
   
   The first column contains the UIDs of the Load Evaluators as seen in DSView.
   Locate the row that contains the UID noted; in this case Row3.
   Locate the value for “nodeid” and make a note of it. In this case it’s 55153.

4. Execute the following SQL statement containing the second context ID and examine the results.
   
   
   The first column contains the UIDs of the Load Evaluators as seen in DSView.
   Locate the row that contains the UID noted; in this case Row3.
   Locate the value for “nodeid” and make a note of it. In this case it is 55154.

5. Execute the following statement to view the records that must be deleted.
   Note: The collected nodeids have been declared and assigned to variables nodeid1 and nodeid2.
   
   
   
   Since the results illustrate the process with nodeid1, the SQL table, and index table, this step can be repeated for both nodeid1 and nodeid2 variables as well as the data table and key table SQL tables.

6. After verifying that the information queried is correct from the data store, run another TSQL script to delete the data.
   Note: The commit statement has been commented out. This is a precautionary measure to ensure that the desired data is deleted.
   
   
   

7. To verify if the corrupted load evaluator is affected by the preceding delete statement, repeat the select queries of Step 5. These results must return no results. After the successful delete statement, run a commit command or a rollback command and all the data remains unchanged.

8. After the corrupted load evaluator has been removed, the Access Management Console can be used to edit, create, delete, and assign load evaluators. All servers that were assigned to the corrupted load evaluator must to be reassigned to an existing or new load evaluator.