The following text lists error messages you might encounter while accessing the database on the ClearPath MCP server and lists possible corrective actions.
A Java Exception has occurred
The Relational Design Center or the Query Design Center has been launched on a systems with multiple Java versions. To resolve the exception, update the existing path to the location of the correct version of Java, as follows:
-
From Control Panel, click System.
-
Click Advanced system settings.
The System Properties dialog box appears.
-
Click Environment Variables....
The Environment Variables dialog box appears.
-
From the System variables scroll box, select Path.
-
Under the System variables scroll box, click Edit....
The Edit environment variable dialog box appears.
-
Click New.
-
Type the installation paths of both the Java Runtime Environment (JRE) and the Java Software Development Kit (JDK). For example,
C:\Program Files\Java\<JRE version>\bin
C:\Program Files\Java\<JDK version>\bin
-
Click OK.
-
On the command prompt, type java-version to verify the current version of Java.
Database is currently being accessed by another MCPSQL Admin process or the job was restarted with a different command.
This message appears for two situations:
-
If a MCPSQL Admin process is started on an enterprise database server that is currently being accessed by a DMSQL Admin process or another MCPSQL Admin process.
-
If a MCPSQL Admin process job is restarted on the same enterprise server but with a different command.
java.net.ConnectException: Connection refused: connect
The specified port number might be incorrect.
Execute the following command on the ClearPath MCP server to identify the correct port number:
NA MCPSQL STATUS
java.lang.OutOfMemoryError:Java
This message appears if there is not enough memory to complete a memory-intensive task, such as copying a large result set in QDC.
To resolve this issue, do the following to increase the Java heap size from the default setting of 64m:
-
Navigate to the directory where the MCPSQL utilities are installed.
-
Open the qdc.cmd file.
-
In the qdc.cmd file, add an -Xmx switch setting to specify the maximum heap size.
For example, if the file contains the following
START /B javaw.exe -jar "C:\Program Files (x86)\Unisys\MCP\MCPSQLUtilities\qdc-60.0.6.jar"
modify it by adding an -Xmx switch setting, as follows:
START /B javaw.exe -Xmx128m -jar "C:\Program Files (x86)\Unisys\MCP\MCPSQLUtilities\qdc-60.0.6.jar"
Note: It is recommended that you set the Java starting heap size (-Xms) to the same value as the maximum heap size to avoid re-sizing the heap during operations. -
Save the qdc.cmd file.
The maximum Java heap size is increased to 128 MB.
Note: If you run out of memory after updating the maximum Java heap size, you can increase the size that you specified in the -Xmx switch setting.
java.net.UnknownHostException: eccsxx
The specified server name is incorrect. Identify the correct name of the server.
No message; you get no response.
The server task SYSTEM/MCPSQL/WORKER might be waiting.
Perform the following steps to respond to this situation:
-
Log on to CANDE using the usercode with which you were trying to connect.
-
Enter the following command to determine whether the Query Processor system files are installed on the secondary family:
FAM*
-
If the secondary family is not in the proper location, change the family statement in the USERDATAFILE.
-
If the secondary family is not an issue, check for waiting entries to try to resolve the problem.
-
Contact the Unisys Support Center if you cannot resolve this problem.
Another possible cause for this error condition is that the SYSTEM/MCPSQL/CONFIG file is waiting on a "NO FILE" condition for the Query Processor configuration file (MCPSQL/CONFIG).
Run the *SYSTEM/MCPSQL/CONFIG program from CANDE to determine the title of the file for which the worker is looking and to verify the syntax of the statements in the file.
Only One MAPPER (DMSIIMAPPER OR DMSIISUPPORT) is Allowed
This message appears for two situations:
-
If two Relational Query Processor users are accessing the same Enterprise Database Server database, but they are using two different MAPPER code files. The default MAPPER code file for MCPSQL is SYSTEM/MCPSQL/DMSIIMAPPER. This can occur if there are multiple MAPPER code files on the system and due to the way the MCP accesses files based on USERCODE family specifications.
Relational Query Processor requires that all users of a particular database must use the same (identical) MAPPER code file. The first Relational Query Processor user to open a database determines the MAPPER code file that must be used by all other Relational Query Processor users of that database.
If possible, remove any unneeded copies of the MAPPER code file and place the MAPPER code file in one location so all users can access that file.
-
If a SQL Query Processor user (DMSQL) and a Relational Query Processor user (MCPSQL) access the same database at the same time. Run only one of these query processor products against an enterprise database server at one time.
Structure <structure name> not mapped. Structure contains an item with DATAENCRYPT option set to true
While performing a SQLVIEW, warning messages are displayed for every structure in the database that contain an item with Field Level Encryption (FLE). Only structures that do not contain an item with FLE are mapped in a SQLVIEW.
The control file for your database cannot be found.
Use the Relational Design Center to update your database.
The MCPSQL client software is incompatible with the MCPSQL server software. Make sure your MCPSQL client and server software are the same product level
The version of Relational Design Center (RDC) installed is not compatible with the MCPSQL server software. Install a version of MCPSQLUtilities that matches the server software level.
The requested operation failed because a record lock deadlock occurred.
An error message might appear when an MCPSQL client attempts to update a record that is read locked by another MCPSQL client.
The following scenarios might cause this error message to appear:
-
Client A sets autocommit to off, and the isolation level is Repeatable Read or Read Committed.
-
Client A performs a SELECT statement, which causes a read lock on the selected records. For example,
select * from mytable where myrec=123;
-
Client B sets autocommit to off, and the isolation level is Repeatable Read or Read Committed.
-
Client B performs an update statement on a record that is locked by Client A. For example,
update mytable set name = 'Alan' where myrec=123;
Client B is blocked waiting for the record to be released.
-
Client A performs an update statement on a record previously locked by Client A. For example,
update mytable set name = 'John' where myrec=123;
-
Client B fails with the deadlock message.
To resolve this issue, set the isolation level of both clients to Serializable.
The total number of data items, parameters, and literals in one query exceeds 1820
This error indicates that the query exceeds the maximum number of allowed operands. An operand includes column names, data values, literals, and table names. For example, for a table with 1024 columns, a simple INSERT statement includes one table name, 1024 column names, and 1024 values for a total of 2049 operands.
| Note: | If column names are not included in the statement, they are implied, and the parser inserts all 1024 columns. |
To resolve this issue, reduce the number of columns in your table to less than 50% of the maximum number of allowed operands. For a simple INSERT statement, the limit on a basic table is 909 columns. Unisys recommends that you normalize your database to keep table sizes as small as possible.
Unable to connect to data source
The user-specified resource or database name is incorrect. Use the Query Design Center to check the catalog of available resources.
Unknown authorization name
This message appears if an incorrect <auth id> or owner is specified for a table or column in a SQL statement.
In releases prior to MCPSQL 60.0, SQL statements used the database name as the <auth ID> or owner in the table name (in the format owner.table). For example, in SELECT * FROM MYDB.product, MYDB is the database name and product is a table.
With the MCPSQL 60.0 release, <auth id> or owner are now specified by the user in the OWNER option of the SQLVIEW command. Or, if the OWNER option was not specified, the user that created the SQLVIEW.
| Note: | For more information about the owner of a database, see Permissions—Assigning Access Rights to Database Structures. |
To resolve this message, modify the SQL statement to use the owner as defined in Permissions—Assigning Access Rights to Database Structures) for <auth ID>. For example, where the owner of database MYDB is USER1 with a table named product:
Select * from USER1.product;
User Limit Exceeded for <database name>
This message appears if more than 1024 users are accessing the same SQL or SQLVIEW database.
Waiting on schema change lock
This message appears if the Schema Administrator program is attempting to replace the existing in-use schema with a newly created schema; this process was likely delayed by a client query that was running for a long period of time.
You can use the NA MCPSQL QUIT NOW command to force remote clients to disconnect from the database. Using the NA MCPSQL QUIT NOW command allows the Schema Administrator program to complete the SQLVIEW request.
You can use the NA MCPSQL STATUS ALL command to list all MCPSQL/WORKERS and examine the impact of the NA MCPSQL QUIT NOW command. Entries that show a usercode and network ID are active database connections that will be terminated by issuing the NA MCPSQL QUIT NOW command.
You should also examine the response from the <database mix number> SM STATUS MIX INTRAN command. The response might show MCPSQL worker tasks that have been in a transaction state for a significant period of time. Such tasks might be the root cause of the “Waiting on schema change lock” error message, and you should consider terminating these tasks before disconnection all clients with the NA MCPSQL QUIT NOW command.
