| Sun Mainframe Administration Tool User's Guide |     |
| Sun Mainframe Administration Tool User's Guide | |
| C H A P T E R 11 |
|
Advanced Tab |
The Advanced tab contains information that is used primarily by customer support personnel. The drop-down list enables you to view the following information types:
The Asynchronous Starts window displays statistics about the region's start queue. It displays the following datapoints:
|
Identifies the request type for this CICS-type message. See CICS Request Type. |
|
|
Identifies the facility that originated this request (TASK, DEST). See Facility. |
|
|
Indicates if the transaction is recoverable. See Recoverable. |
|
|
Remote terminal ID assigned to this request. See Remote Terminal ID. |
|
|
Remote transaction ID assigned to this request. See Remote Transaction ID. |
|
|
System ID assigned to this request. See Remote System ID. |
|
|
Request ID assigned to this request. See Request ID. |
|
|
Timestamp when this entry was placed in the start queue. Not implemented in this release. |
|
|
Indicates the transaction start code (S, QD, SD). See Start Code. |
|
|
Displays the current status of an entry on the start queue. See Status. |
|
|
Indicates the exec interface block (EIB) task number that submitted this request. See Submitting Task. |
|
|
Terminal ID assigned to this request. See Terminal ID. |
|
|
Requested transaction ID. See Transaction ID. |
|
|
Indicates the item number of the temporary storage queue. See TS Queue Item. |
|
|
Identifies the queue assigned to this request. See TS Queue ID. |
|
|
Indicates the type of entry on the start queue. The only valid value is CICS, which represents a CICS statement. |
The Facility datapoint shows the origin of the asynchronously started transaction.
The value of the datapoint is based on the facility type. Facility can be either the transaction ID of the CICS request or the name of the transient data (TD) queue that triggered the request.
Use the Start Code datapoint to determine if the started transaction was requested by another transaction or triggered from a TD queue. When the Start Code datapoint is S or SD, it indicates that the CICS request came from another transaction. When the Start Code datapoint is QD, the request was triggered from a TD queue.
The TS Queue Item datapoint displays the number of the item within a temporary storage (TS) queue that was passed with a START request.
The value of the datapoint is an integer number that identifies a specific record in a TS queue.
Sun MTP passes data specified with the FROM parameter to the START statement to the requested transaction by means of the TS Queue Item in a particular TS queue. The TS queue can be specified by the user using the QUEUE parameter to the START command, or it is generated by Sun MTP. The item number of the data within the TS queue accompanies the entry so that Sun MTP can retrieve the appropriate data.
The TS Queue ID datapoint contains the name of a TS queue specified as the QUEUE parameter to the EXEC CICS START command.
The value of the datapoint is a one- to eight-character queue name.
A particular item in the TS queue holds the data specified by the FROM parameter to the START command. The TS Queue Item datapoint provides a means to pass data along with the START request.
The value of the Recoverable datapoint indicates if the asynchronously started transaction is recoverable or not.
The values of the datapoint can be either true or false.
If the Recoverable value is false, Sun MTP submits the transaction immediately, even for a rollback. However, if the Recoverable value is true, Sun MTP does not submit the transaction when there is a rollback. In this sense, the transaction is protected.
The Remote Terminal ID datapoint contains the value specified by the RTERMID parameter of the EXEC CICS START command. It represents the terminal name against which the transaction will be run on the remote system.
The value of the datapoint is a one- to four-character name.
The remote terminal name is passed to the asynchronously started transaction, which receives the name by issuing a RETRIEVE command with the RTERMID option.
The Remote Transaction ID datapoint represents the value specified by the RTRANSID parameter in the EXEC CICS START command. It is the remote transaction identifier to be used on the remote system when routing this transaction entry to the remote system.
The value of the datapoint is a one- to four-character name.
The transaction that issues the START statement specifies the remote transaction ID. The issued transaction receives the remote transaction ID by issuing a RETRIEVE command with the RTRANSID option.
The RTRANSID parameter contains any data the starting transaction wants to pass to the asynchronously started transaction.
The CICS Request Type datapoint shows the type of CICS command request of an entry on the start queue.
The value of the CICS Request Type is:
START: Start request (EXEC CICS START)
If there is a value in the CICS Request Type datapoint, it indicates that the entry on the start queue originated from a CICS application program. The CICS Request Type datapoint contains a value only when the Type datapoint contains the value CICS.
The Request ID datapoint displays the value of the REQID parameter specified in the EXEC CICS START command.
The value of the datapoint can be a one- to eight-character name.
The value of the Request ID datapoint uniquely identifies the associated START statement. This value can be used by a CANCEL command to cancel the transaction before it executes.
If the Request ID value is specified for the start queue entry, it indicates that the transaction can be cancelled prior to being executed.
The Start Code datapoint shows the start code of a START request.
The Start Code datapoint indicates how a request was started (by means of a START command or a transient data trigger) and whether additional data was passed along with the request. When a transaction issues a START request (identified by Task#), the requestor can use the FROM parameter to the START command to specify data to pass along with the request. The transaction requested by the START statement obtains this data by means of the RETREIVE command.
The Status datapoint displays the current status of an entry on the start queue.
|
More entries on the process queue than there are transaction servers |
|
|
Connection is not defined in the terminal table system entries |
|
The Status value of the entry changes based on Sun MTP region resources. If a resource is not available to process the entry, the appropriate value is set for the entry. When the resource is available, the status is reset. Values marked with an asterisk (*) are not resolvable, and will not relinquish their start queue slots.
If a transaction remains in the start queue for a long time, the Status datapoint indicates exactly why the transaction is waiting. As long as the Status datapoint is blank, there is nothing blocking the transaction's eventual execution.
The Remote System ID datapoint contains the value specified by the SYSID parameter in the EXEC CICS START command. This is the name of the remote system where the transaction is to run.
The Remote System ID value can be any valid one- to four-character remote system name specified in the system entries portion of the terminal table.
The Remote System ID value can be used to determine the number of jobs that are currently waiting in the start queue that are to be executed on another system.
The Submitting Task datapoint displays the exec interface block (EIB) task number of the transaction that submitted a transaction to the start queue.
The value of the Submitting Task datapoint is an integer number for a specific instance of a transaction.
The Terminal ID datapoint contains the value specified by the TERMID parameter in the EXEC CICS START command. It is the terminal identifier against which the transaction is to be started.
The value of the Terminal ID datapoint is any valid four-character name. If the START command is to execute locally, the Terminal ID value must match an entry in the terminal table.
If the Terminal ID datapoint is omitted, the task will run without a terminal, which means that the application program cannot perform I/O to a terminal.
The Transaction ID datapoint displays the value specified by the TRANSID parameter in an EXEC CICS START command. The Transaction ID is the transaction identifier.
The value of the Transaction ID datapoint can be any valid four-character transaction identifier specified in the region's transaction table.
The Transaction ID datapoint represents the transaction that was initiated by the end user at the terminal, by a another transaction by means of the RETURN statement, or by a transient data queue trigger.
The Transaction ID values on this window give you a good idea of the types of transactions that are waiting to be executed.
The Environment Variables pane displays each environment variable set for the region and its value.
The System Gates panel displays information related to Sun MTP gating primitives. This information is useful for investigating system throughput. The following datapoints are displayed:
|
Identifies the gate. See Name. |
|
|
Indicates the total time, in seconds, spent attempting to obtain ownership of the gate. This includes any time waiting for the gate due to an existing owner. |
|
|
Indicates the delta value of gate waits. See Delta Waits. |
|
|
Indicates the maximum time, in seconds, that a requestor was forced to wait for a gate. |
|
|
Indicates the maximum time, in seconds, that an owner held the gate. |
|
|
Indicates the maximum number of requestors forced to wait for this gate. |
|
|
Indicates the total amount of time, in seconds, that all owners held this gate. |
|
|
Indicates the number of times this gate has been owned since the region was started. This value is initialized to zero and region startup and is incremented each time the gate is locked. |
|
|
Indicates the total number of times a process waited for this gate. See Total Waits |
|
|
Indicates the current number of requestors waiting for this gate. |
The Name datapoint shows the name of a system gate used by a region. The server processes of a region employ these gates to control access to shared resources.
The following table describes each gate:
A gate is a mechanism that allows multiple processes to gain access to a resource in a mutually exclusive manner. A Sun MTP region's server processes may use all of the gates described in TABLE 11-3. The only user-configurable gates are the buffer and index gates, whose quantity is determined by the number of buffers specified in the VCT. Monitor the usage of the KXSEMBUF0-f gates to adjust the number of buffers for optimum performance. When the numbers of waits on the buffer gates keep climbing, increase the number of buffers to spread out the gate access to the data and indexes in the buffers.
The Delta Waits datapoint shows the number of times server processes have waited on a particular gate during the last polling interval.
The Delta Waits datapoint is an integer value that is calculated every polling interval.
This datapoint should be monitored when the system seems to be running slowly. A large number of Delta Waits for an extended period of time can indicate a bottleneck in the system due to a particular gate.
The Total Waits datapoint indicates the total number of times a requestor was forced to wait for the gate.
The datapoint value is a number that is initialized to zero at region startup. Sun MTP increments the number of Total Waits for a particular gate every time a requestor has to wait to obtain the gate.
The Total Waits datapoint indicates how much contention a gate has. Two gates may have large numbers of locks, but one may have more waits than the other.
The System Queues window displays details about the Sun MTP interprocess communications (IPC) message queues.
|
Note - The operating system maintains these queues, not the Sun MTP region. |
The following datapoints are displayed on this window:
|
Identifies the queue. See Name. |
|
|
Indicates the maximum queue size in bytes. See Maximum Size. |
|
|
Indicates the total number of bytes currently in use by this queue. See Current Size. |
|
|
Indicates the number of messages currently in this queue. See Current Messages. |
|
|
Process ID of process that issued the last msgrcv operation. |
|
|
Timestamp of last msgrcv operation. See Last Receive Time |
|
|
Process ID of process that issued the last msgsnd operation. |
|
|
Indicates the maximum number of bytes used at one time. See Peak Size. |
|
|
Indicates the maximum number of messages queued at one time. See Peak Messages. |
|
|
Indicates the total number of bytes processed by this queue. |
|
|
Indicates the total number of messages processed by this queue. |
The Name datapoint shows the name of a system queue used by Sun MTP. A queue is an operating system resource enabling interprocess communication. Sun MTP creates a number of queues that are used to pass messages and requests between the various components and server processes that make up a region.
The possible values for the Name datapoint are: KXPRTQ, KXPROCQ, KXRECOVQ, KXSTRTQ, KXTERMQ, KXCLASSQnnn, and KXTRANQnnn.
The following table describes each type of queue:
The most important queues relating to system performance are KXPROCQ, KXCLASSQ, and KXRECOVQ.
The Maximum Size datapoint represents the maximum number of bytes a system queue can hold at any one time. A queue is an operating system resource enabling interprocess communication. This value is part of your operating system configuration (kernel parameter).
This numeric value is obtained directly from the operating system, not from the region.
If, for some reason, the capacity of a queue is reached, the process trying to put a message or request on the queue must wait until another process takes a message or request off the queue. This condition can cause region throughput to slow down. You can detect this condition by comparing the Peak Size datapoint to the corresponding Maximum Size datapoint. If the Peak Size datapoint reaches the allowed size of the queue, the queue size must be increased.
Refer to your operating system's system administrator's guide for information about increasing the maximum size of a system queue.
Monitor the Maximum Size datapoint, especially during the initial deployment of a Sun MTP region, to help determine an appropriate queue size. Resizing system queues not only requires shutting down the region, but the operating system as well, affecting all activity on the machine.
The Current Size datapoint represents the number of bytes currently on a particular system queue. A queue is an operating system resource enabling interprocess communication.
The value of the Current Size datapoint is a numeric quantity obtained directly from the operating system.
The value of the Current Size datapoint indicates how heavily a queue is used. The value is the total number of bytes used by messages on a queue. The number of messages appears in the Current Messages datapoint.
The Current Size value is a snapshot in time: at each refresh interval, the administration tool queries the operating system for the size of the queue.
When the value of the Peak Size datapoint approaches the maximum size for the process queue, the transaction class queue, or any of the transaction server queues, more transaction servers can be configured to alleviate server overload for the region. Examine the Performance screens to see how the transaction servers are being utilized. If transactions are always being run in the transaction servers during peak production hours and messages are always waiting to be processed on the process queue or transaction class queues, increase the number of transaction servers.
The Current Messages datapoint represents the number of messages or requests currently on a particular system queue. A queue is an operating system resource enabling interprocess communication.
The value of the Current Messages datapoint is a numeric quantity obtained directly from the operating system.
A message on a queue consists of a certain number of bytes. Each message on a particular queue is the same size, but each queue can have its own message size. To determine the message size for a queue, divide the number of Current Size by the number of Current Messages. Dividing the Maximum Size by the size of a message tells you the effective maximum number of messages allowed on a particular queue.
The value of Current Messages on the process queue and transaction class queues tells how many requests are waiting to be processed. If this number remains high as response time diminishes, consider configuring more transaction servers to accommodate the message processing.
The value of Current Messages on the recovery queue indicates the number of requests (before image, writes, or rollbacks) that are waiting to be processed.
The Last Receive Time datapoint shows the last time a process took an entry from a system queue. A queue is an operating system resource enabling interprocess communication.
The Last Receive Time datapoint has the following format: hh:mm:ss. This value is obtained directly from the operating system.
This statistic can be used to determine if processes are taking messages or requests from the queue in a timely manner.
Consider a situation where the Current Size value reaches the allowed size. This indicates that the queue membership has reached its capacity in terms of size. A normal course of action could be to increase the size of the queue or adjust the number of processes that read the queue. However, the Last Receive Time datapoint might reveal that processes are not taking messages off the queue for some reason, such as system problems, aborted processes, looping processes, and so on. Make sure to evaluate all possible causes before changing your configuration.
The Peak Size datapoint represents the maximum number of bytes that have been recorded on a particular system queue. A queue is an operating system resource enabling interprocess communication.
The value of the Peak Size datapoint is a numeric quantity obtained by comparing the current value to the value of Current Size and setting Peak Size to the higher of the two.
This value indicates the queue usage (along with the Peak Messages datapoint) at peak load in production processing. Note that the actual peak usage of a queue might occur between polling intervals and could escape the administration tool's inquiries to the operating system.
If the value of the Peak Size datapoint approaches the value of the Maximum Size datapoint, consider increasing the size of the queue.
The Peak Messages datapoint indicates the maximum number of messages that have been recorded on a particular system queue. A queue is an operating system resource enabling interprocess communication.
The value of the Peak Messages datapoint is a numeric quantity obtained by comparing the current value to the value of Current Messages and setting Peak Messages to be the maximum of the two.
This value indicates the queue usage (along with the Peak Size datapoint) at peak loads in production processing. The actual peak usage of a queue might occur between polling intervals and may escape the administration tool's inquiries to the operating system.
Most systems have a kernel parameter that specifies the maximum number of messages that can be on a queue at any given time. This parameter is similar to the Maximum Size datapoint of a queue (which is measured in bytes). Consult your operating system's system administrator's guide to find out where to look for this value. If the Peak Messages datapoint approaches this value, it must be increased to obtain better performance from the region.
The size of the messages put on the queue can also affect the maximum number of messages allowed on a queue, because of a queue's maximum size. For example, if a queue has an allowed size of 100 bytes and a maximum of 10 messages, and the messages for this queue are 50 bytes each, the queue has an effective limit of two messages.
| Sun Mainframe Administration Tool User's Guide | 817-0572-11 | |
Copyright © 2004, Sun Microsystems, Inc. All rights reserved.