18.1.6 Facility for changing the process-down message when a transaction is cancelled

Organization of this subsection
(1) Overview
(2) Function

(1) Overview

If a server process is terminated by a forced termination request due to an interrupt at the client during transaction execution, or if a server process connected to a UAP is terminated by the pdcancel command, HiRDB displays the KFPS01820-E and KFPO00105-E messages that indicate termination of the server process. These messages are also displayed when a server process is terminated due to some abnormality in the server's processing. To determine the cause of such a termination, you must also check other messages, such as KFPS00993-I, that are displayed by HiRDB.

Use of the facility for changing the process-down message when a transaction is cancelled enables you to change the IDs of the KFPS01820-E and KFPO00105-E messages that are displayed in the case of client-related operations and errors. When you do this, you can easily identify a termination of server processing that was caused by a client-related operation or error simply by checking the message IDs (eliminating the need to check other messages).

(2) Function

This facility changes only the message IDs; it does not change the message text or output information (the original message text and output information are still displayed as is even after the message ID has been changed).

(a) Messages subject to change

Table 18-5 lists the messages that can be changed by this facility.

Table 18-5 Messages whose message IDs can be changed

Message ID that can be changedDescription of the messageMessage ID after change
KFPS01820-EServer process was terminated.KFPS01852-W
KFPO00105-EServer process was terminated because an error occurred (abort message).KFPO00115-W

For details about these messages, see the manual HiRDB Version 8 Messages.

(b) Whether or not message IDs are changed

Table 18-6 shows whether or not message IDs are changed, depending on the process-down event.

Table 18-6 Whether or not message IDs are changed depending on the process-down event

No.Proc-down eventCause of process-downMsg ID chgdRemarks
User opEnv setClient errorClient comm errServ errorS-to-S comm error
1Forced termination request due to an interrupt at the clientC----------YesHiRDB server process was terminated during execution of a transaction due to a forced termination request caused by an interrupt at the client.#1
2HiRDB server process killed due to a client process-down----CC----YesFor a client process using XA, see No. 11.
3HiRDB server process killed because PDCWAITTIME was exceeded--C#2--CC--NoHiRDB cannot identify the cause.#3
4Process-down (exit) at HiRDB server because PDSWAITTIME was exceeded--CCC----Yes--
5Process-down (exit) at HiRDB server because PDSWATCHTIME was exceeded--CCC----Yes--
6HiRDB server process killed due to forced termination of unit, unit down, or transaction recovery during system switchoverC--------#4--#4YesApplicable to another unit's branch recovery. This is applicable only to a HiRDB/Parallel Server.
7HiRDB server process killed due to forced termination of unit, unit down, or utility branch recovery during system switchoverC--------#4--#4YesApplicable not only to recovery in the local unit but also to another unit's branch recovery.
8HiRDB server killed by pdcancel command (including a killed utility)C----#4----#4--#4Yes--
9HiRDB server process killed due to forced termination of transaction by pdfgt command (transaction branch recovery)C----#4----#4--#4Yes--
10HiRDB server process killed by the facility for monitoring free area for system log fileC#5C#6--------YesKFPS01160-E message is displayed.
11HiRDB server process killed by XA completion instruction issued by the transaction manager in an extension of transaction recoveryC------CCNoHiRDB cannot identify the cause.#7
12HiRDB server process killed due to recovery when the synchronization skip count exceeded the boundary valueC#8C#6----C--NoHiRDB cannot determine the cause.
13Process-down (exit) at HiRDB server due to timeout during communication between HiRDB server processes--------CCNo--
14Communication request processing from a shutdown unit when recovery- unnecessary front-end server is used--------CCNo--
15Process-down (exit) at server due to transaction completion error when recovery- unnecessary front-end server is used--------CCNo--
16Process-down at HiRDB server when automatic log unloading facility was terminated due to a server failure--------CCNo--
17HiRDB server process failure shutdown (such as internal failure, process failure, or transaction start error)--------CCNo--
18HiRDB server process killed during transaction and utility branch recovery due to HiRDB server process failure shutdown----------#4--#4YesRecovery processing caused by No. 17.
Legend:
Proc-down event: Process-down event
User op: User operation
Env set: Environment settings
Client comm err: Client communication error
Serv error: Server error
S-to-S comm error: Server-to-server communication error
Msg ID chgd: Message ID changed
C: Considered to be the cause of the process-down.
--: Not the cause of the process-down.
Yes: Message is changed.
No: Message is not changed.
#1
This is applicable to UAPs that use DBPARTNER or DABroker to access HiRDB, and to UAPs that access HiRDB via ODBC. This also includes cases such as when the applicable UAP is terminated by pressing the Ctrl + C keys on the keyboard.
#2
Applicable when the value is less than the transaction processing time.
#3
This requires user judgment because HiRDB cannot determine whether this is a timeout caused by environment settings or by a server failure. For this reason, message output is required.
#4
Although this is not an error event, it can cause an error event.
#5
Applicable when the volume of transactions exceeds the expected value.
#6
Applicable when a boundary value is invalid.
#7
This is caused by a completion instruction from the transaction manager. If a transaction is cancelled due to a client failure, by the user, or due to a server failure, the transaction manager issues a completion instruction. In this case, HiRDB cannot identify the cause of the completion instruction.
#8
Applicable when the issued SQL statement is invalid.
(c) Unneeded operands

To use this facility, either omit the pd_cancel_down_msgchange operand or specify Y.

If you have specified v6compatible or v7compatible in the pd_sysdef_default_option operand, the default value of the pd_cancel_down_msgchange operand is set to N; therefore, you must specify Y in the pd_cancel_down_msgchange operand.