Tmax 5 SP2 Fix#5

This chapter briefly describes newly added and changed features and bug fixes introduced in Tmax 5 SP2 Fix #5. For detailed information, refer to the relevant guide.

1. Added features

This section describes the major features added.

1.1. Engine

  • Support for FD_32768 field size in Tmax 5 SP2

    • Tmax 5 SP2 supports both the existing FD_16384 and the new FD_32768 field size.

    • The default remains FD_16384. The FD_32768 version is available as a separate porting release upon request.

    • Related issue: IMS 300065

1.2. Client/Server

  • Configurable socket buffer size between server and CLH

    • Added support for the CLOPT="-D socket buffer size" option in the SERVER section.

    • This option sets the socket buffer size for communication between CLH and the server, in bytes. The default value is 16384 (16KB). If not set, this value is automatically applied.

      If the size of the message the server is attempting to transmit exceeds the socket buffer size, CLH must repeatedly read the data up to the configured buffer size. This can cause delays in send() calls as the server’s send buffer slowly empties. This delay can be mitigated by adjusting the socket buffer size using the -D option. However, the actual socket buffer size may vary depending on the operating system.

    • Usage examples

      *SERVER
svr1 SVGNAME = svg1, MIN =1, MAX=1, CLOPT = "-D 102400"

    • Related issue: IMS 246419

  • Added TCP_NODELAY option for client-to-CLH communication

    • Previously, TCP_NODELAY could only be set in WinTmaxStart(). It can now also be set in tpstart().

    • When calling tpstart(), the TCP_NODELAY option is applied by setting the environment variable TMAX_CLH_TCP_OPTION=YES.

    • This helps reduce delays when sending messages from clients to CLH.

    • Related issue: IMS 320733

1.3. Management tools

  • Added slog recording for management commands

    • When executing the management command in tmadmin, success and failure results are now recorded in slog.

    • By recording the main operation commands of the Tmax system, such as suspend and resume, in slog, command history is automatically recorded, and this enhances traceability and simplifies system management.

    • Usage examples

      $$1 tmax (tmadm): sp -v svr2
svr(svr2) is suspended
$$2 tmax (tmadm): rs -v noserver
no such svr(noserver)

(slog)
ADM.268121.150731:(I) ADM4000 ADM_SUSPEND Success command(sp -v svr2) [ADMC001]
ADM.268121.150752:(I) ADM4000 ADM_RESUME Failure command(rs -v noserver) [ADMC000]

    • Related issue: IMS 290748

1.4. Utility

  • Added svcrpt second/minute output and changed hourly output format

    • svcrpt -n h|m|s outputs in hours/minutes/seconds. (Default: h)

    • The existing svcrpt was output horizontally at one-hour intervals, but it has been changed to a vertical output format.

    • Usage examples

      $ svcrpt -i svclog.07032025 -n s

Services issued
        between Thu Jul  3 17:37:30 2025
        and     Thu Jul  3 17:37:37 2025

TOUPPER

           Time            Num            Avg
   ------------   ------------   ------------
       17:37:30              1       0.000164
       17:37:33              2       0.000035
       17:37:34              1       0.000012
       17:37:36              1       0.000058
   ------------   ------------   ------------
          total              5       0.000061

    • Related issue: IMS 270266

  • Added the -a option to svcrpt for continuous output

    • When svcrpt -a is used, the Time, Num, and Avg values are output at fixed intervals, even when the service is not called.

    • This enables consistent performance data collection in Excel, making it easier to visualize service trends.

    • Usage examples

      $ svcrpt -i svclog.07032025 -n s -a

Services issued
        between Thu Jul  3 17:37:30 2025
        and     Thu Jul  3 17:37:37 2025

TOUPPER

           Time            Num            Avg
   ------------   ------------   ------------
       17:37:30              1       0.000164
       17:37:31              0       0.000000
       17:37:32              0       0.000000
       17:37:33              2       0.000035
       17:37:34              1       0.000012
       17:37:35              0       0.000000
       17:37:36              1       0.000058
   ------------   ------------   ------------
          total              5       0.000061

    • Related issue: IMS 290775

  • Added validation of LOAD setting for COUSIN groups in cfl command

    • A function was added to check whether the LOAD value of each server group set to COUSIN in cfl complies with the rules.

    • Validation rules

      LOAD value Rule

      > 0

      All server groups must have LOAD >= 1

      0

      All server groups must have LOAD = 0

      -1

      All server groups must have LOAD = -1

      -2

      All server groups must have LOAD = -2 or -3

    • Usage examples

      $ cfl -i 341298.m
(E) CFL3202 LOAD mismatch: svg1 has LOAD=-2, but COUSIN gw13 has LOAD=1 [CFL0314]

    • Related issue: IMS 341298

1.5. TDL

  • Added the tdlcall(TDL_RTLD_GLOBAL) flag function for accessing external variables during transaction execution

    • Setting the TDL_RTLD_GLOBAL flag in the flags of tdlcall*() allows loaded external symbols to be referenced from other modules. This can be used when accessing External variables is needed in the OpenFrame OSC environment.

    • Related issue: IMS 272557

  • Added the -e option to return an error code if tdlupdate is performed on only some nodes during tdlupdate

    • When using the tdlupdate -e option in a multi-node (MODE=SHARED) environment, if tdlupdate fails on some nodes, an error code is returned. If not set, tdlupdate returns a success return value (0) on the master node even if it fails on some nodes.

    • Usage examples

      $ tdlupdate -m func001 -e
TDLUPDATE.redhat73.14389.174419:(E) TDLUTIL0028 socket connect error: nodename = node2, racport = 3478 [TDL0018]
TDLUPDATE.redhat73.14389.174419:(I) TDLUTIL0095 /home/tmaxha/tmax/run/func001.64d5f4e3 is being processed [TDL0417]
TDLUPDATE.redhat73.14389.174419:(I) TDLUTIL0118 Synchronization succeeded [TDL0064]
TDLUPDATE.redhat73.14389.174419:(I) TDLUTIL0083 func001.64d5f4e3 updated successfully [TDL0443]
$ echo $?
9

    • Related issue: IMS 272557

  • Added the tdlcall/tdlload execution time measurement function

    • Setting the environment variable TDL_ENABLE_DL_TIMER=Y measures the execution time for dlopen, dlsym, and module calls when calling tdlcall*, tdlload* functions in the debug libraries (libsvrd, libclid, libtdlcalld). (Default: N)

    • Example debug log

      53937546.182625:(D1) SVR [DLOPEN][/home/tmaxha/tmax/run/tdlcall_A.680760c1] START:182625.800097, END:182625.800838, ELAP:0.000741
53937546.182625:(D1) SVR [DLSYM][tdlcall_A] START:182625.800872, END:182625.800875, ELAP:0.000003
53937546.182625:(D1) SVR [FUNC][tdlcall_A] START:182625.800949, END:182625.800950, ELAP:0.000001

    • Related issue:https://ims.tmaxsoft.com/tody/ims/issue/issueView.do?issueId=339865[IMS 339865], 171058

1.6. HostLink

  • Added the HostLink module functionality on Linux

    • Supports the HostLink module on Linux.

    • Related issue: IMS 274149

1.7. Others

  • Provides header file change history and version functionality

    • Added a function to record version information in the header file provided by Tmax.

    • The change history of each header file is provided in the changelogs subdirectory in the format <header file name>.log, and changes after the release are additionally recorded by revision.

    • The target directories are /cobinc, /topinc, /tuxinc, /usrinc, /usrinc/tmaxrpc, and some files from these are excluded.

    • Example header file

      /* -------------------------- usrinc/atmi.h -------------------------- */
/*                                                                     */
/*             Copyright (c) 2000 Tmax Soft Co., Ltd                   */
/*                      All Rights Reserved                            */
/*             Version: 5.0 SP #2 Fix #5 r11882 (final_release)        */
/*             Change Logs: usrinc/changelogs/atmi.h.log               */
/*                                                                     */
/* ------------------------------------------------------------------- */

    • Example change log file

      /* ------------------ usrinc/changelogs/atmi.h.log ------------------- */
/*                                                                     */
/*              Copyright (c) 2000 Tmax Soft Co., Ltd                  */
/*                       All Rights Reserved                           */
/*                                                                     */
/* ------------------------------------------------------------------- */

[5.0 SP #2 Fix #5 r11882 (final_release)] - 2025-07-17
  * Fixed:
  * Added: Add header version to comment
  * Change:

    • Related issue: IMS 283360

2. Changed features

This section describes the major changes to the features.

2.1. Header file changes

This section describes changes to the structure definition and function signature in the header file. Modules using APIs that are related to changed items may require recompilation to ensure compatibility with existing binaries.

Between the Tmax 5 SP2 Fix4 and Tmax 5 SP2 Fix5 releases, there are no changes to the definition or function signature.

2.2. Engine

  • Separated the TSKILLTIMEOUT setting into TSINITKILLTIMEOUT and TSDONEKILLTIMEOUT

    • Previously, the timeouts for tpsvrinit() and tpsvrdone() were set to be the same using TSKILLTIMEOUT, but this has been improved to allow the startup and shutdown times to be set separately.

    • TSINITKILLTIMEOUT specifies the wait time until the server is terminated when a hang occurs while executing tpsvrinit().

    • TSDONEKILLTIMEOUT specifies the wait time until the server is terminated when a hang occurs while executing tpsvrdone().

    • TSKILLTIMEOUT has been removed, so if you are using the old settings, you must check.

    • Related issue: IMS 342294

  • Changed TSM to not support multi-threaded server types

    • The method of TSM (tmapm) terminating a thread or process when a thread’s service times out in the STD_MT server type has been removed, and has been changed so that it no longer works in the STD_MT type.

    • Related issue: IMS 261392

  • Improved server restart prevention function according to TSDONEKILLTIMEOUT when tmdown -s/-p option is terminated

    • When shutting down the server with the tmdown -s or -p option, there was a problem where the server would automatically restart if TSDONEKILLTIMEOUT was triggered and the server was shut down.

    • On the other hand, the tmdown -S, -A, -g options were processed to terminate without rebooting in the same situation.

    • This fix improves consistency of behavior so that tmdown -s/-p options terminate normally without rebooting.

    • Related issue: IMS 338658

2.3. Management tools

  • Improved *count item output method of ti, cfg -d, cfg -n

    • The svrcount entry of cfg -d/-n has been changed to output only the actual number of servers, excluding the MAXRQSVR and MAXGWSVR counts.

    • The svg/svr count in cfg -d has been changed to output the sum of all nodes rather than the value for each node.

    • Changed so that svccount entries are additionally output when using cfg -d.

    • Related issue: IMS 281571

  • Improved Server and TMS count output function when reopen_db is used

    • When entering the xa server group name, the number of servers and TMS belonging to the server group is displayed.

      $$1 tmax1 (tmadm): reopen_db svg_ora
DB reopen request sent (svrcnt : 1, tmscnt : 2)

    • Related issue: IMS 280635

  • Improved the order of commands when using help so that they are sorted alphabetically

    • When executing the help command, the list is modified to be sorted alphabetically.

    • Related issue: IMS 289533

2.4. TDL

  • Improved delay when running tdlupdate in a multi-node environment (SHARED mode)

    • The tdlupdate -N option synchronizes only the target modules in a multi-node (MODE=SHARED) environment, reducing the execution time compared to a full synchronization. If this option is not specified, all modules registered in SRUNDIR are synchronized with LRUNDIR, which can be time-consuming.

    • However, for stable synchronization in a multi-node (MODE=SHARED) environment, it is recommended not to specify the -N option.

    • If you are experiencing synchronization-related issues or service stability is critical, we recommend choosing a time slot with low load and performing a full synchronization using the tdlupdate -c option.

    • Related issue: IMS 309269

  • Unified the usage of tdlshm/tdlseqno -m/-M options

    • All tdl utilities now reflect the -m/-M options consistently.

      • -m: library name (library)

      • -M: function name (function)

    • The existing option behavior of tdlshm and tdlseqno was changed from [-m module|-M library] to [-m library|-M function].

    • Note that depending on the option change, it may behave differently than expected when used in the same way as before.

    • Related issue: IMS 336543

2.5. TCP gateway

  • Improved TCPGW abnormal termination phenomenon due to excessive header size definition in custom.h

    • When TCPGW processes a request from a remote node, an exception handling is applied to recalculate the buffer size for reading the header according to msg_header_size, thereby improving the abnormal termination phenomenon caused by excessive header size.

    • However, the msg_header_t structure should only define fields at the message metadata level and should not contain actual data (payload).

    • Related issue: IMS 335870

2.6. Java Gateway

  • JEUSGW CLOPT -D option debug log improvement

    • Improved clarity of the meaning of the debug logs being output.

      • index[%d] → xa : clh cpc index[%d]

      • nonxa get clhi : index[%d] → nonxa : clh cpc index[%d]

      • reply_to_clh → reply_to_clh ind[%d]

    • When outputting debug logs with the CLOPT -D option, the phrase 'JEUSGW' is now printed before the log message.

    • Since the log output with the CLOPT -D option is also a debug log, it was changed to be output as (D#) type instead of (I) type.

    • The CLOPT -D option has been modified to insert a space between the module name and the log contents in the log.

    • An example debug log is as follows:

      23376.175628:(D1) GATEWAY JEUSGW xa : clh cpc index[1]
23376.175628:(D1) GATEWAY JEUSGW reply_to_clh ind[1]
23376.175628:(D1) GATEWAY JEUSGW request_from_clh ind[1] msgtype[1003] svc[FDLINS]

    • Related issue: IMS 294110

2.7. Others

  • Optional modules are provided to improve installation efficiency.

    • Some less frequently used modules have been excluded from the default installer to optimize installation capacity and deployment efficiency.

      • WebService Gateway (23): tmmbfgen*, xwsdlgen*, untmmbfgen*, wsgw*, libexpatMT*, libhttpgw*, libwsgw*

      • SNMP Agent (2): tmsnmpd*

      • Tmax RPC (22): idlc*, librpccli*, librpcsvr*, librpcsql*, librpcsyb*, librpctbr*

      • WebAdmin (2): twagent*

    • If the module is required, it can be provided separately upon request through IMS.

  • Modified tuxinc/atmi.h to include usrlinc/atmi.h

    • Fixed a compilation error that occurred due to the include order of the atmi.h header file in the Tuxedo conversion AP source.

    • Related issue: IMS 133211

  • OSS zlib version upgrade (1.1.4 → 1.3.1)

    • The version of the zlib compression library was upgraded from 1.1.4 to 1.3.1.

    • Related issue: IMS 343222

3. Bug Patch

This section describes the bug fixes.

3.1. Engine

  • When CLHOPT -x 1 is set, '.CHK_CONN' is not output to slog when calling tmaxchk_conn (positive number)

    • When CLHOPT -x 1 is set, '.CHK_CONN' is not printed in slog when calling tmaxchk_conn (positive number).

    • Related issue: IMS 265012

  • Fixed the issue where TSM did not check for timeouts for services scheduled on the second and subsequent channels when cpc was 2 or more

    • Fixed the issue where TSM (tmapm) did not check service timeout for services scheduled on the second or higher CPC channel when CPC is 2 or more.

    • Related issue: IMS 261392

  • Fixed the issue of mksvr server service not being registered when configuring COUSIN SVG

    • Fixed the issue where the service of a server built with mksvr was not properly registered in an environment where two SVGs were configured as cousins on one node.

    • Related issue: IMS 274905

  • Fixed the FUNCNAME non-functioning issue in the service section

    • Fixed the issue where setting FUNCNAME in the service section did not work.

    • Related issue: IMS 279126

  • Fixed the issue where some logs were lost when svclog changed the date

    • svclog has fixed the issue of some logs being lost when the date is changed.

    • Related issue: IMS 272941

  • Fixed the issue where SHMKEY+5 (TSM) permission was not applied to IPCPERM

    • The phenomenon where SHMKEY+5(TSM) was always generated as 700 and IPCPERM was not applied has been fixed.

    • Related issue: IMS 279126

  • Fixed the TPENOREADY error that occurs when requesting remote domain services from a node without COUSIN settings

    • The phenomenon of a TPENOREADY error occurring when calling a remote domain service from a node not bound to COUSIN after the remote domain is terminated and restarted in the COUSIN environment of the domain gateway has been fixed.

    • Related issue: IMS 299306

  • Fixed memory leak and unnecessary server information transmission when using the CLHOPT -c option

    • The phenomenon of the CLH process’s memory continuously increasing and causing abnormal termination even when there is no load when using the CLHOPT -c option has been fixed.

    • Additionally, the phenomenon of information from servers that do not belong to the cousin being transmitted has been corrected, and only the necessary server information is transmitted.

    • Related issue: IMS 291229

  • Improved connection loss and duplicate error when starting multiple multi-nodes simultaneously

    • The phenomenon of intermittent connection loss or duplicate errors occurring when starting multiple multi-nodes simultaneously has been improved.

    • In a multi-node environment, each node may attempt to connect to other nodes simultaneously, which may result in connection failures or errors due to connection requests colliding or being mistimed.

      The more nodes you have, the more likely you are to encounter this issue, so we recommend the following startup method:

      • Integrated startup using racd on the management node

      • When manually starting nodes, they are started sequentially with a set time interval between each node.

    • Related issue: IMS 294570

  • Fixed CLH abnormal termination when UCS, TCPGW, DOMGW, and TMS servers are down on Linux k2.6/k4.18 platforms

    • Fixed the issue where the CLH process would terminate abnormally when the UCS, TCPGW, DOMGW, and TMS servers went down on Linux k2.6 and k4.18 kernel-based platforms.

    • Related issue:https://ims.tmaxsoft.com/tody/ims/issue/issueView.do?issueId=323891[IMS 323891], 327765, 339452

  • Fixed the issue where IRT was not applied when the JEUS_ASYNC Gateway configured with IRT in Domain2 entered through Domain1

    • Configuration: client(tpcall) → dom1(node1/2) (TMAXNONTX) → dom2(node3/4) (JEUS_ASYNC) → JEUS(1/2)

    • When the JEUS_ASYNC Gateway configured on node3 of dom2 is down, a problem occurred where requests succeeded/failed 1:1.

    • The issue where IRT was not applied when JEUS_ASYNC Gateway configured with IRT in dom2 entered TMAXNONTX Gateway through dom1 has been fixed.

    • Related issue: IMS 328821

  • Fixed the issue where TMM would generate SIGFPE when restarting after CLH and server abnormal termination

    • Fixed the issue where SIGFPE would occur if the server process registered with TMM before the CLH during a restart after the CLH and server processes were abnormally terminated.

    • Related issue: IMS 337778

  • Fixed the issue where core would crash due to failure to detach tmm’s shared memory (shm) during tmdown

    • Fixed the issue where a core would occur if the tmm process failed to release shared memory (detach) when performing tmdown.

    • Related issue: IMS 342144

  • Fixed the issue where slog files were not created when a license warning log occurred in tmm

    • Fixed the issue where a new date slog file was not created when a Cloud license period warning log (TMM3073, TMM3074) was output from tmm.

    • Related issue: IMS 344270

3.2. Client/Server

  • When outputting tmadmin API node config, change the number of svg and svr to the actual number of server groups and servers

    • When outputting node config information in the tmadmin API, the number of svg and svr has been modified to be output as the number of server groups and servers of the actual node.

    • Related issue: IMS 281571

  • Fixed the issue where data is transferred to a different channel when calling tpalloc - tpforward when using JEUS_ASYNC GW

    • Removed the problem that occurs in the tpacall → JEUS_ASYNC → IPPR(tpforward) → SVC(tpreturn) call flow.

    • When a request message arrives from JEUS_ASYNC, the remote channel information is stored in the header, and the corresponding channel information is read and the message is transmitted when responding.

    • When tpalloc() is called in the IPPR service, a new header and message are created, but when tpforward() is called, the remote channel information stored in the existing header is not copied to the new header, so there was a problem in which the response from JEUS_ASYNC was sent to a different channel.

    • This fix resolves the issue.

    • Related issue: IMS 267365

  • Fixed server abnormal termination issue when applying libsvr to previous Tmax (tmm) after Tmax 5 SP2 Fix4 release

    • The issue of abnormal server termination when applying libsvr version later than Tmax 5 SP2 Fix4 r11811 to Tmax (tmm) version prior to the release has been fixed.

    • Related issue: IMS 268117

  • Fixed a crash when calling the STD_MT server type

    • The phenomenon of rows being hung when receiving a response from CLH after calling a service of the STD_MT server type if the seqno is different has been fixed.

    • Related issue: IMS 260033

  • Fixed a startup error on remote node server when using tpadvertise() on multi-node

    • After patch r11820, the phenomenon of a startup error occurring on the remote node server when using tpadvertise() in multi-node has been fixed.

    • Related issue: IMS 274905

  • Fixed the issue where log file date was not updated when using the CDATE macro in the STD_MT server type

    • When using the CDATE macro in the STD_MT server type, the issue where a new log file is not created even after the date is changed and log files from the previous date continue to be written has been fixed.

    • Related issue: IMS 280404

  • Fixed the issue where certain servers would not shut down when using tmdown -R

    • Fixed the issue where servers using the tpacall() - tpgetrply() pattern would intermittently not terminate and would hang when using the tmdown -R command.

    • Related issue: IMS 305133

  • Fixed the TPEINVAL issue when using tpgetcliinfo() in a multi-clh environment

    • During the application of FD 32768 patch (r11840), the clid format was changed, which caused a conflict between libsvr`, which uses the changed format, and the Tmax system, which uses the existing format, resulting in a TPEINVAL error when calling tpgetcliinfo().

    • The error was prevented by modifying the clid and spri formats to be dynamically set according to the FD_SETSIZE value.

    • Related issue: IMS 308336

  • Fixed the issue where a response was not received after tprelay() of UCS when requesting from JEUS_ASYNC Gateway

    • Configuration: tpacall → JEUS_ASYNC → IPPR(tpforward) → A(tpforward) → UCS(tprelay) → B(tpreturn)

    • The JEUS_ASYNC Gateway stores the channel information used in a request in the header. However, when UCS calls tprelay() and recreates the header message, this channel information is not updated, which causes a problem.

    • As a result, even if a response comes from UCS, the JEUS_ASYNC Gateway cannot know which channel the request was made through, so a problem occurred where the response was sent through the wrong channel.

    • To address this issue, we’ve modified tprelay() to update the request channel information when calling it. Now, the JEUS_ASYNC Gateway can receive responses on the correct channel.

    • Related issue: IMS 342332

  • Fixed the issue where the server would not shut down when executing the tmadmin API in the server process during tmdown -S processing

    • During tmdown -S processing, some server processes requested tmadmin(TMADM_SVC_STAT) due to a timing issue, the tmdown request arrived first, and tmadmin(TMADM_SVC_STAT) received an unexpected msgtype. This caused a TPESYSTEM error, and then the tmadmin API response message was passed to the server process, and the (E) SVR0209 internal error: invalid msg type 1124 [SVR0034] error log was printed. There was a problem that the server process did not terminate.

    • We have modified the service to process messages normally even when a message type different from the requested message type is received during service processing, so that problems do not occur.

    • Related issue: IMS 318730

  • Fixed the SVR3012 error that occurs when repeatedly restarting the STD_MT type server

    • Fixed the issue where SVR3012 error occurred when repeatedly restarting the STD_MT type server.

    • Related issue: IMS 333991

3.3. Management tools

  • Fixed an error in outputting avg value for STD_MT server type st -s/-p

    • The phenomenon of the avg value of st -s/-p being abnormally output when calling a service of the STD_MT server type has been fixed.

    • Related issue: IMS 260033

  • Fixed the ci -P output error when there are 2 or more CLHs

    • When there are two or more CLHs and there are empty slots among the entire CLHs, the phenomenon of qpcount, emcount, ippaddr port and tmax_port being incorrectly output when executing the tmadmin ci -P command has been fixed.

    • Related issue: IMS 270291

  • Fixed the issue where the MAXRQCPC entry was missing when executing the *cfg -n command

    • The missing maxrqcpc value when executing tmadmin’s cfg -n command has been modified to be output normally.

    • Related issue: IMS 288064

  • Fixed the issue when entering an incorrect server group name in the *reopen_db command

    • Fixed the issue where tmadmin would terminate abnormally when a non-existent server group name was entered.

    • Fixed the issue where an unrelated error message was displayed when entering a non-xa server group name.

    • Added missing commands reopen_db, notify_reconnect_clh, and admnoti to tmadmin help.

    • Related issue: IMS 280635

  • Improved stat command usage messages and error handling

    • Reflected missing options and modifications in the usage message of the stat command.

      • -pod, -j options added.

      • The output format ([], {}, |) has been modified to be consistent.

    • Usage messages are now displayed when non-existent options such as -a and -h are entered.

    • Fixed sorting-related error handling for the -s and -p options.

    • Related issue: IMS 290421

  • Fixed the ADM0211 error when registering a cfgadd dynamic service

    • There was a problem where the error (W) ADM0211 General Information: invalid cfgfile in SERVICE, KAQM1024 is omitted [ADM0434] occurred when dynamically registering a service with the cfgadd command. This occurred when the service index value became tangled when there were a large number of registered services.

    • It has been modified so that errors do not occur even in situations like the above.

    • Related issue: IMS 312610

  • Fixed abnormal termination when querying tms information with st -t on HP-UX

    • Fixed the issue where abnormal termination occurred when retrieving tms information with the st -t command when MAXGW is set to an odd number and GW exists in an HP-UX environment.

    • Related issue: IMS 314055

  • Fixed an error that occurred for service requests after the *rbs (-S/-s) command

    • Fixed the issue where the server would reboot infinitely without processing the service request after rebooting the POD server (MIN=0) with rbs -S.

    • Fixed the issue where replacing a binary with rbs -S while processing a service on a TCS server with MIN >= 1 would result in scheduling only to the newly started server process.

    • Fixed the issue where service requests would not be scheduled when suspending/resumeing a service with sp/rs -s SVC.

    • When requesting RBS to a POD server that is processing a service, the existing service is processed normally, but the issue where subsequent service requests are not processed and are queued has been fixed.

    • Fixed the issue where the server would not restart for the MIN number of times when the server was terminated by the LIFESPAN option while executing RBS.

    • Related issue: IMS 341369

3.4. Utility

  • Fixed the minimum value error in the CFL error message when setting the server-side CPC=0

    • When CPC=0 is set in the server section, the error message displayed when executing CFL incorrectly displayed the minimum value of the valid range as 0. This has been corrected to correctly display it as 1.

    • Related issue: IMS 288067

  • Fixed abnormal termination when the number of services exceeds 1024 when running svcrpt

    • After r11823, the issue of abnormal termination when the number of registered services exceeds 1024 when executing svcrpt has been fixed.

    • Note: r11823 added second/minute output capability to svcrpt and changed the hourly output format.

    • Related issue: IMS 306722

  • Fixed an error where libraries could not be found when using the -t STD_MT option when building a dynamic server using mksvr

    • Fixed an error that caused the library to not be found during compilation when using the -t STD_MT option when building a dynamic server using mksvr.

    • Related issue: IMS 339556

  • Fixed the issue where _fdl was included even when the -h option was specified when using fdlc

    • Since version 5.0.2.4.r11405, when using fdlc and specifying the -h option, _fdl is not included and a header file is created with the specified file name.

    • Note: In r11405, the -k option was added, which improved the functionality to generate class or interface names exactly as the Java file names specified with the -h option.

    • Related issue: IMS 341900

  • Fixed the issue where *TS*KILLTIMEOUT was not applied when redefined

    • The phenomenon in which TIMEOUT is applied as the value set in the DEFAULT section without being disabled even when TS*KILLTIMEOUT is redefined to 0 in each lower server setting after setting the TS*KILLTIMEOUT option in the DEFAULT section has been fixed.

    • Related issue: IMS 337491

  • Fixed the TS*KILLTIMEOUT Timeout error that occurred in the TIP server process

    • The phenomenon in which the TS*KILLTIMEOUT timeout was applied to server processes (TIPs) that do not have tpsvrinit() or tpsvrdone() has been fixed.

    • Related issue: IMS 337491

3.5. TDL

  • Fixed the issue where some options were missing in tdlupdate/tdlshm Usage

    • Added -e, -N, -h, -l file, -r tdldir options that were missing from the Usage output of the tdlupdate command.

    • Added -V, -b, -B bakfile, and -a options that were missing from the Usage output of the tdlshm command.

    • Related issue: IMS 343608

  • Corrected tdlclean help content error and -m/-M option description

    • The incorrectly written [-m module|-M library] in the Usage options of tdlclean has been corrected to [-m library|-M function].

    • Consistently reflected the description of the -m/-M options for all tdl utilities.

      • -m: library name (library)

      • -M: function name (function)

    • Related issue: IMS 336543

3.6. TCP gateway

  • Fixed core occurrence when the uid value modified in put_msg_info() is larger than the uid max value (2,097,152) that divides the data structure

    • In put_msg_info(), if the modified uid value is greater than the uid max value (2,097,152) that divides the data structure, a TCPGW0057 error is generated.

      (E) SVR6226 UID(50002) range is exceeded in put_msg_info, Valid range = (MAX : 50001) [TCPGW0057]
(E) SVR6226 UID(2097153) range is exceeded in put_msg_info, Valid range = (MAX : 2097152) [TCPGW0057]

    • Related issue: IMS 280893

3.7. Java Gateway

  • Fixed the issue where an abnormal termination occurred during the response reception process when using JEUS GW

    • The phenomenon of GW being abnormally terminated when the XID Seqno of an XA inbound request coming into JEUS GW is negative has been fixed.

    • Related issue: IMS 287383

3.8. HostLink

  • Fixed an error when calling cimsgw for the first time immediately after startup

    • The phenomenon in which an incorrect response message was transmitted when the cimsgw service was first called immediately after startup, resulting in an error, and normal operation was fixed for subsequent calls of the same service.

    • Related issue: IMS 278406

3.9. Others

  • Fixed incorrect comment in tmadm_svc_conf_body in tmadmin.h

    • Fixed the issue where the comment for the tmadm_svc_conf_body structure in the tmadmin.h header file was incorrectly marked as TMADM_SVR_CONF instead of TMADM_SVC_CONF.

    • Related issue: IMS 283325