Sybase Technical Library - Product Manuals Home
[Search Forms] [Previous Section with Hits] [Next Section with Hits] [Clear Search] Expand Search

Chapter 8 User Exits [Table of Contents] Appendix A Error Messages

Installation and Administration Guide for Mainframe Connect for DB2 MVS-CICS

[-] Chapter 9 Troubleshooting

Chapter 9

Troubleshooting

This chapter provides guidelines for what to do when a client application cannot access data on the mainframe. The guidelines include how to locate an error and cover the most frequent problems.

This chapter also includes operational and process flow overviews for Enterprise Connect. The information applies to both CICS LU 6.2 and TCP/IP environments, except where differentiated.

The sections in this chapter are:

This chapter describes mainframe troubleshooting. For comparable DirectConnect information, see the Transaction Router Service User's Guide for DirectConnect and the Access Service User's Guide for DirectConnect for MVS.

Verifying connectivity after installing MainframeConnect for DB2/MVS-CICS

Be sure the AMD2 transaction processor and the client and server connections are configured correctly. You can use one of the standalone utilities (snaping or cicsping) shipped with Open ServerConnect to test the network connections between the workstation and mainframe. Use snaping for Open ServerConnect for LU 6.2 and cicsping for Open ServerConnect for TCP/IP.

These utilities do not require DirectConnect to be active. For details on using them, see the Installation Guide for DirectConnect for MVS.

If these utilities fail to operate correctly, the failure is caused by a configuration error or line outage problem. Based on the error message returned, you can narrow your search. Resolution may require coordination with the DirectConnect administrator to make changes to the configuration.

After snaping or cicsping runs successfully, be sure the DirectConnect administrator defined all the necessary remote procedure call (RPC) and connection information to DirectConnect. Be sure that DirectConnect is started. Use the guidelines in the following section for further problem determination.

Where to start troubleshooting

System administrators at the client, DirectConnect server and mainframe must check components systematically to locate the source of the problem. Check for problems in this sequence:

For any of these problems, the appropriate system administrator should:

Common problems and suggested solutions

Problems often can be traced to configuration errors or network, line, modem, or adapter outages.

Configuration errors

Configuration errors are most often the cause of communications failure. To resolve these errors, you need the following information, which is created when the network is installed and successfully implemented:

Verify that this information is still the same as before the error occurred. If it is not the same, determine if a recent change is contributing to the problem. The following subsections explain the causes and actions for common configuration errors.

Cannot establish a session

Cause

Any of the following:

Action

Correct the spelling.

Coordinate with the DirectConnect administrator to check connection and modename profiles using the snaping or cicsping utility shipped with the product.

Session established, but the transaction does not run

Cause

Any of the following:

Action

Verify definitions.

Coordinate with the DirectConnect administrator for correct security and transaction ID setups.

SDLC line or token ring is not up

Cause

Address incorrectly configured with NCP (assumes correct line or modem setup)

Action

Check both ends of the SDLC station or Token Ring address configuration.

SDLC link and PU are active, but the LU Is not active

Cause

Any of the following:

Action

Use the SDLC trace and error log facilities to find the error.

Mainframe network operational failure

On the mainframe, there are two frequent causes of operational errors:

In these cases, either:

When you are contacted about such messages, reactivate the necessary mainframe resources.

Network session or line failures

Line, adapter, or modem outages result in error messages at both the SNA console and DirectConnect. DirectConnect records the message and, if possible, sends a similar error message to any affected clients.

Preventative measures

Intermittent hardware errors and line degradation problems can disrupt processing, but can be difficult to find. You should check periodically for these types of problems. For example:

Components

MainframeConnect for DB2/MVS-CICS support consists of several components on the IBM mainframe and the DirectConnect platform, as Figure 9-1 shows. These components provide tracing and logging, which you can use to locate errors. This section includes examples of tracing and logging for CICS.

The following figure shows Open ServerConnect and DirectConnect support components in a CICS environment:

Figure 9-1: Open ServerConnect and DirectConnect support componentsraster

DirectConnect support

DirectConnect performs the following functions:

Each TRS or access service includes a unique server name, which clients use to select a server for communication. Each TRS or access service has its own set of configuration information.

As shown in Figure 9-1 , DirectConnect uses four files:

For more information, see the Transaction Router Service User's Guide for DirectConnect and the Access Service User's Guide for DirectConnect for MVS.

DirectConnect communications with MainframeConnect for DB2/MVS-CICS

DirectConnect depends on the communications support, supplied by the platform vendor, to communicate with MainframeConnect for DB2/MVS-CICS. Depending on the server-based communications software installed, DirectConnect uses SNA LU 6.2 or TCP/IP Uplink support.

SNA LU 6.2

SNA software uses the SNA trace file to record SDLC/SNA traffic between the server and mainframe. The vendor trace utility extracts this file.

For Windows NT platforms, the error log file records errors that SNA Server for Windows NT detects.

TCP/IP uplink

For CICS TCP/IP environments, third-party TCP/IP trace facilities provide a way of obtaining low-level TCP/IP level traces between the mainframe and DirectConnect. For AIX platforms, the error log file records errors that SNA Server/6000 detects. The IBM error log report utility extracts this information. Microsoft SNA Server for Windows NT also has a trace process.

Mainframe Communications with DirectConnect

Mainframe-based communications support provides the "transport" level of function. Depending on the mainframe communications software installed, Gateway-Library uses SNA/NCP or TCP/IP for MVS.

SNA/NCP

For CICS LU 6.2 environments, you can use the SNA General Trace Facility (GTF) files to trace SDLC/SNA traffic between Gateway-Library and the mainframe. The IBM TAP utility extracts this information.

TCP/IP for MVS

For CICS TCP/IP environments, you can use the Netstat facility to check the status of TCP/IP connections, as well as to deactivate them if problems occur.

You can use the TCP/IP trace facility to trace traffic between DirectConnect and the mainframe.

Sybase TCP/IP Listener

The Sybase listener transaction provides the interface between CICS and MVS/TCP. The listener enables CICS applications to use the following TCP/IP communication facilities:

The Sybase listener TRACE parameter and Open ServerConnect DEBUGSW parameter settings provide CICS user exit traces, denoting each TCP call issued and the associated return call.

CICS

For CICS, Gateway-Library tracing stores information on the TDS traffic between the mainframe and server in the VSAM ESDS file, SYTDLOG1. This information includes any errors detected in the traffic.

Open Server's Trace flag will write Open Server TCP/IP Trace point to the CICS Message User log, as well as the CICS Aux trace facility.

Some TD calls fill up internal TDS buffers before sending them out to the network. For example, a TDSNDROW or TDSNDMSG call does not cause execution of a corresponding CICS EXEC SEND call, unless the TDS buffer becomes full.

Warning!

To avoid losing records, periodically archive or delete the trace records on SYTDLOG1. Trace records append to this file until it is full, then it rejects any further records.

MainframeConnect for DB2/MVS-CICS

MainframeConnect for DB2/MVS-CICS provides the AMD2 transaction, which automatically processes client SQL language requests using the DB2 dynamic SQL facilities.

Using this product, client applications can communicate directly with DirectConnect or with another server that communicates with DirectConnect, such as Adaptive Server or ASE/CIS (as shown in the following figure):

Figure 9-2: MainframeConnect for DB2/MVS-CICS and DirectConnect support componentsrasterTracing

MainframeConnect for DB2/MVS-CICS has a trace flag that will write out MFC trace points to a CICS Temp Storage Queue name "CE + 'first 6 bytes of the user id'."

For MainframeConnect for DB2/MVS-CICS, you can also use the standard DirectConnect tracing support. See Figure 9-1 .

Logging

You can determine the errors logged. Enter changes in SOURCE library member AMD2MAMX, as detailed in MainframeConnect for DB2/MVS-CICS installation instructions. For example, to indicate that you want a message logged, set LOG=Y.

SYRTMAMX TYPE=ENTRY, MSGNR=32000, SQLCODE=0, LEVEL=11, STATEC=0, STATESC=0, LOG=Y, SQLFATAL=N

Warning!

If a connection fails, the error log is the only method of tracking an error. To prevent the log from running out of room, initialize the log file regularly.

In a CICS environment, MainframeConnect for DB2/MVS-CICS provides the AMD2LOG and AMD2LOGO files. The AMD2 transaction writes errors to AMD2LOG, including task ID, term ID, timestamp, and the unique message code for each error. AMD2 also returns an error message to the client.

Under CICS, if AMD2LOG is full or no longer accepts messages, AMD2 writes errors to the overflow file, AMD2LOGO. AMD2 then issues error message 33219:

MainframeConnect for DB2/MVS-CICS error (Write to log file AMD2LOG failed, switching to overflow log AMD2LOGO).

Under CICS, you can close, archive (if desired), and purge AMD2LOG when it becomes full. During the time AMD2LOG is closed, AMD2 writes messages to AMD2LOGO that you can close, archive, and purge once you reopen AMD2LOG.

In this example, AMD2 writes an internal error 32000 to AMD2LOG (or AMD2LOGO, if AMD2LOG is full).

For CICS environments, you must define the error log files to CICS in an FCT entry as follows:

Coordination

System administrators at the mainframe, DirectConnect and client need to coordinate troubleshooting efforts. To aid in your analysis, this section describes the processing flow from the client through DirectConnect to the mainframe.

Processing flow and DirectConnect requirements

DirectConnect processing can involve CT-Library, DB-Library, and ODBC. The following figure shows DirectConnect processing with DB-Library (for more information, see the DirectConnect for MVS Access Service User's Guide for DirectConnect for MVS):

Figure 9-3: Client-to-DirectConnect-to-Mainframe Processingraster

The following points describe the sequence shown in Figure 9-3 and highlight Enterprise Connect requirements:

The following points describe both RPC and dynamic SQL processing flows.

AMD2 supports user-defined transactions, which are comparable to long-running transactions based on Gateway-Library support.

Process flow during attention sequences

Any of the following actions results in an attention sequence:

When an attention sequence is issued, the process flow is as follows:


Parse exit routine [Table of Contents] Appendix A Error Messages