Tunning message queues IPC parameters

The following table shows the parameters to be set in relation to the message queue parameters:

Parameter	Description	Recommended Value
msgmnb	Maximum number of bytes that can be stored in a queue	Calculate the maximum number of bytes per queue from the number of concurrent transactions and the message size. Take into account that Tuxedo starts writing message to a file when queue reaches 75% capacity.
msgmni	Maximum number of queues in the system	tmboot-c command indicates the recommended number to be used for the current configuration. For reference, this value must correspond approximately to: MAXACCESSERS (number of servers and native clients) + response queue Servers - Server mode sets MSSQ + Total MSSQ + 7 (where 7 is the approximate number of queues Tuxedo administrative servers.)
msgtql	Maximum number of concurrent IPC messages in the system	Calculate the number of concurrent messages based on the number of concurrent transactions.

Tuning shared memory IPC parameters

The following table shows the parameters to be set in relation to the shared memory parameters:

Parameter	Description	Recommended Value
shmmax	Maximum size in bytes of a shared memory segment	This value multiplied by shmseg must be sufficient to accommodate the BB (Bulletin Board). tmboot -c command indicates the size required for the BB
shmmni	Maximum number of identifiers of shared memory segments in the system	Tuxedo needs a segment for the BBL, for each queue space and WSL table.
shmseg	Maximum number of shared memory segments per process	This value multiplied by shmmax should be sufficient to accommodate the BB (Bulletin Board).

Tuning semaphore IPC parameters

The following table shows the parameters to be set in relation to the semaphore parameters:

Parameter	Description	Recommended Value
semmni	Maximum number of active semaphore set	Using tmboot-c command to determine the appropriate value
semmns	Maximum number of semaphores in the system	Set at least equal to MAXACCESSERS - MAXWSCLIENTS +13 (approximate number of semaphores for Tuxedo administrative servers). Using tmboot-c command to determine the appropriate value
semmnu	Number of undo structures in the system.	Since undo structure is required for each process accessing the Bulletin Board, this value must equal SEMMNS Using tmboot-c command to determine the appropriate value
semmsl	Maximum number of semaphores per semaphore set	SEMMNS / SEMMNI Using tmboot-c command to determine the appropriate value
semume	Maximum number of undo entries per undo structure.	Equal to SEMMNS Using tmboot-c command to determine the appropriate value

Oracle Tuxedo at Oracle Open World 2012

Following is listing of Tuxedo related activities at OOW 2012. These activities are primarily focus on Tuxedo12c, launched in August 2012.

 

Activity	Date / Time	Name
Conference Sessions	Mon 1 Oct, 2012	10:45 AM - 11:45 AM, Oracle Tuxedo: What’s New in 12 c, Strategy, and Roadmap, Moscone South - 309
Conference Sessions	Mon 1 Oct, 2012	4:45 PM - 5:45 PM, Simplify Operations, Administration, and Management of Oracle Tuxedo Applications, Marriott Marquis - Golden Gate C3
Conference Sessions	Wed 3 Oct, 2012	3:30 PM - 4:30 PM, The Art and Practice of Mainframe Migration and Modernization, Moscone South - 309
Conference Sessions	Thu 4 Oct, 2012	2:15 PM - 3:15 PM, High-Performance, Scalable Enterprise Messaging for C/C++/COBOL Applications, Marriott Marquis - Salon 7
HOL (Hands-on Lab)	Tues 2 Oct, 2012	5:00 PM - 6:00 PM Deploy, Manage, and Monitor Oracle Tuxedo Applications in the Enterprise Cloud, Marriott Marquis - Salon ¾
HOL (Hands-on Lab)	Wed 3 Oct, 2012	1:15 PM - 2:15 PM, Develop C/C++ Applications for the Cloud with Oracle Tuxedo and Oracle Solaris Studio, Marriott Marquis - Salon 5/6
BOF (Birds-of-a-Feather)	Mon 1 Oct, 2012	6:15 PM - 7:00 PM, Develop Scalable, Highly Available Enterprise Services in Java with Oracle Tuxedo, Marriott Marquis - Golden Gate C1
Demos		Oracle Tuxedo: #1 Enterprise Cloud Platform for C/C++/COBOL Apps,  Moscone South, Right - S-215
Demos		Mainframe Rehost with Oracle Tuxedo Runtimes for CICS, IMS, and Batch, Moscone South, Right - S-216
Tuxedo Customer Appreciation Dinner	Monday 1 Oct, 2012	7:30 PM - Please contact your Oracle Account Representative to attend. Limited seating.

Parameters related to the size of the Bulletin Board

It is advisable to keep the size of the Bulletin Board as tight as possible to the real burden of our application, without forget to keep additional space to support new servers started dynamically to meet peak demands. The values ​​that affect the size of BB are:

Parameter	Description
MAXSERVERS	Maximum number of servers in the domain, including administrative servers such as BBLS and TMS.
MAXSERVICES	Maximum number of services that can be published in the domain.
MAXGTT	Maximum number of transaction supported by Tuxedo simultaneously.
MAXCONV	Maximum number of simultaneous conversations in a machine
MAXWSCLIENTS	Maximum number of remote clients that can connect to a machine Tuxedo domain

Transactions per second in mission critical applications Tuxedo

According to case studies and ORACLE benchmarking testing, the following are the measure benchmarks:

• 100,000 transactions per second (tps) with a response time less than the second.

• 13.5 K systems tps for credit card authorization system.

• 56K tps on billing systems in telecommunication companies.

Monitor Tuxedo Application using TXRPT utility

The txrpt statistical tool is used to measure the services performance. Below a brief guide shows the use of the tool.

• txrpt tool works with trace service requests stored in files. In those files, Tuxedo stores the start and end date / time of every service request done.
• This trace generation is enabled using the -r option in the ubbconfig file for each server entry. Tuxedo, by default, writes the trace file in standard error, so it is recommended to redirect the trace log to a separate file for each server, using the -e option. The following entry is an example of a configuration server tracing:

"server1" SRVGRP = "servergroup" srvid = 1CLOPT = "-A -r

-E /tuxedo/traces/server1.trc"

 When using the-r option, ensure that the ULOGDEBUG variable is not initialized or is initialized to 'N', this is because if it is initialized as 'Y', Tuxedo also writes debug information on the standard error file.

• The txrpt tool is used to generate metrics and service average times. These times are presented in centiseconds. The semantics of this is:

txrpt [-t] [-n names] [-d mm / dd] [-s time] [-e time]

Where,

-T: Optional. If specified, indicates that services should be sorted by the total service time in descending way. If it is not specified, the services are ordered by the number of invocations in descending way.

-N: Optional. Restrict output to the services listed separated by comma.
-D mm / dd: Optional. If it is specified, the statistics are generated on the date mentioned, if it is not, statistics are generated for the current day. In any case, the utility generates statistics just for a day.
-S hr [: min [: sec]]: Optional. Restrict the generation of statistics to those entries starting after the appointed time.

The following is an example of use of the tool:

txrpt -d 01/12 -s8:00 -e9:00  < /tuxedo/traces/server1.trc

START AFTER:    Tue, Jan 12, 2012 08:00:00 AM

END BEFORE:     Tue, Nov 29, 2011 09:00:00 AM

      SERVICE SUMMARY REPORT

SVCNAME         8a-9a          TOTALS

                Num/Avg         Num/Avg

--------------- --------        -------

TOUPPER         5/0.5           5/0.5

--------------- -------         -------

TOTALS          5/0.5           5/0.5

As shown, the first lines indicate the range of dates / times used, then displays a list of services and a column for each rank of an hour, in that cell shows the number of service invocations and the average time response. The last column shows total number of invocations and average times.