Supplying specific details to your support organization helps them determine the problem more quickly.
The /usr/lpp/sna/bin/getsnapd command creates a compressed tar file of all the SNA files needed for debugging. The compressed file is pd.tar.Z and is left in the current directory. You should invoke /usr/lpp/sna/bin/getsnapd as soon after you encounter a problem as possible.
The following table lists the items that must be provided to the support organization when you report a problem. This list also shows the name of the file that needs to be moved to the directory sent to the support center.
List of Items, with File Names, to Be Provided to Technical Support | |
Item | File to be Saved |
Problem Definition | README |
SNA Server Profiles | Anyname |
SNA Internal Error Log | /var/sna/snaservice.X |
System Error Log | /var/adm/ras/erriog |
Link Station Trace | /var/sna/ProfName |
The following table lists additional information that may be required as the problem is investigated.
Possible Additional Information for Problem Investigation | |
Item | File to Be Saved |
SNA API Trace | /var/adm/ras/trcfile |
888 Error Codes | |
System Dump | |
LU0 Trace | /var/lu0Sec /var/lu0/LU0Prime |
HCON Trace | hcon.trace |
SNA_ABEND | /vqr/sna/SNA_ABEND |
Using Hook IDs |
Note: You must log in as the root user to run the commands necessary to provide the appropriate information to the support center.
First, and most important, is a clear definition of the problem. This definition should be included in a file called README on the diskette. This file should also include a list of the files on the diskette and a brief description of those files. Be sure to answer the following questions in the problem description:
Always provide your profiles when you have a problem. To copy profiles to the /tmp/pmr directory, log in as the root user and enter the following command:
exportsna -A -U -f /tmp/pmr/xxxxx.profiles
To provide only the session profile and any profiles pointed to by the session profile, enter the following command:
exportsna -s SessName -f /tmp/pmr/yyyyy.profiles
where SessName is the name of your session profile.
The SNA service log is used to keep track of SNA errors. This log is located in directory /var/sna and is named snaservice.X, where X is an integer that is incremented with each start/stop cycle of SNA or when toggled with the sna -setlogs -t command. The snaservice log is always active and is initially set to /var/sna/snaservices.1 It is automatically toggled to the next integer (up to 10 and back to 1) each time SNA is stopped and restarted, toggled using the sna -setlogs -t command, or by using the smit fast path smit _snatracecontrol . To capture the service log, start SNA with the following command:
sna -s
Make sure you check the latest snaservice log using the ls -l /var/sna command and observing the date and time information or before starting SNA you may enter the command:
rm /var/sna/snaservice*
to assure the current service log will be /var/sna/snaservice.1.
Run a test case to recreate the problem.
Stop SNA to close the snaservice log file or stop SNA tracing using the sna -trace off command and then toggle the log file using the sna -setlogs -t command to begin a new snaservice log file. Copy the snaservice log that was current for the duration of the test to the /tmp/pmr directory.
cp /var/sna/snaservice.1 /tmp/pmr
Before updating a system error log, ensure that it is cleared, so that all entries relate to the current problem. The errclear command deletes error log entries that are older than the number of days specified by the Days parameter. To delete all error log entries, specify a 0 value for the Days parameter. Enter the following commands:
errclear 0
Run a test case to reproduce the problem, then enter the following command:
cp /var/adm/ras/errlog /tmp/pmr/errlog
The link station trace can be set to automatically start when the link station is started, by using SMIT or the chsnaobj command. To turn on tracing for an X.25 link station, enter the following command:
chsnaobj -tlink_station -wx.25 -Qyes -wlong PROFILE_NAME
or use the smit fastpath _snaX25attcch to select and change the Trace link? field in the link station profile.
To capture link station trace information for a particular problem, first enable link station tracing for the link station using either the chsnaobj command or through SMIT. Start SNA and the link station and run a test case to recreate the problem. After the problem has been reproduced, turn off link station tracing to close the trace file using the following command:
sna -t off -l PROFILE_NAME
Copy the resulting link station trace file to /tmp/pmr:
cp /var/sna/PROFILE_NAME /tmp/pmr/PROFILE_NAME
The SNA Application Program Interface (API) trace logs the entry and exit points of the SNA API routines, as well as significant events and command details of the operating system and library subroutines for the following application program interfaces:
To enable API traces, the system event trace file must be designated as the SNA Server/6000 service log. The API traces may be enabled by using SMIT or the -trace option of the sna command. For example, the command; sna -trace only -l 'RMTLNK' -a 'on' turns on tracing for the SNA API on link station RMTLNK. For the CPI interface the command would be; sna -trace only -l 'RMTLNK' -c on . For Generic SNA API the command is; sna -trace only -l 'RMTLNK' -g on .
Trace enabling may also be accomplished using SMIT and the snatrace fastpath and setting 'on ' for the appropriate application program interface.
To capture API trace information in the service log, first enable tracing for the appropriate API. Then start SNA and run the application. Next, stop SNA or stop SNA tracing with the command; sna -trace off .
You may need to supply LU 0 information to the support center.
The smit _snaLU0 fast path provides information for LU0 primary logical unit and the LU0 secondary logical unit profiles. The exportsna command output contains the LU0 configuration information.
The lu0 command initiates the LU0 subsystem. The LU0 subsystem initiates and centralizes control of both the LU0 primary and secondary support servers. The servers' data paths are independent of each other. However, pass-through support provides for the logical coupling of the two servers.
The lu0 command provides a common operator interface through the interactive commands. These commands allow you to manipulate the LU0 subsystem while it is running, as well as helping to minimize system resource consumption. You can use the interactive commands to display status summaries, start a secondary server, stop a server, terminate all servers, and exit the program. The lu0server process may also be run in the background by entering the lu0 & command. While running the lu0server in the background, the lu0sndmsg command is used to send commands to the lu0server. A line trace facility for the lu0 subsystem may be initiated with the lu0 -T command or if the lu0server is started in the background, with the lu0sndmsg -T command. The trace facility records the first 20 bytes of the SNA PIU block for traffic going in either direction.
Up to 15 characters of run data is saved in the file. Two files are created containing trace information for the primary server and the secondary server, respectively: /var/lu0/LU0Prime and /var/lu0/LU0Sec.
To create LU0 line trace files, perform the following:
# sna - sna /*start sna # lu0 -T /*start the lu0server process or # lu0 & /*either in the foreground or background # lu0sndmsg -T /*specifying the -T option for trace if /*lu0 is background
Run a test case to reproduce the problem, then stop the lu0server process by entering X to the lu0 shell (foreground) or kill -9 xxxxx (background), where xxxxx is the process ID. The process ID was returned from the lu0 & command. You can also obtain the process ID by entering kill %1 if no other background processing has been started from the shell.
LU0 API tracing is enabled for tracing by selecting yes in the Enable API trace field in the smit _snaLU0plu and smit _snaLU0slu screens for each primary and secondary LU0 profiles. LU0 API tracing may also be enabled by the following commands:
chsnaobj -t lu0_plu -A yes PRIMARY_LU0_PROFILE_NAME
chsnaobj -t lu0_slu -A yes SECONDARY_LU0_PROFILE_NAME
To create LU0 API trace files, perform the following steps:
# sna -s sna /* start sna
# sna -stop sna /* stop sna
The SNA API trace information is contained in the current snaservice log:
# cp /var/sna/snaservice.1 /tmp/pmr/snaservice.1 /*save it
HCON has a trace capability built into the product. To run the trace and copy the trace file, enter the following command:
sna -s sna
Run a test case to reproduce the problem. Then enter the following commands:
sna -stop sna
cp hcon.trace /tmp/pmr/hcon.trace
The SNA_ABEND file is created any time that SNA Server processes an unexpected exit. Stopping SNA Server with the -t cancel flag or the -t force flag also creates this file.
If the /var/sna/SNA_ABEND file is being created when the problem occurs, the file needs to be sent to the support organization. To save the SNA_ABEND file:
sna -s sna
cp /var/sna/SNA_ABEND /tmp/pmr/SNA_ABEND
The support organization may ask you to generate an event trace. All of the AIX traces generate arguments to the trace command. The trace command uses hook IDs to determine what to trace. The support organization needs both a formatted and an unformatted copy of the event trace.
View the /usr/include/sys/trchkid.h file to see a list of all hook IDs.
For example, to create and capture event information for the X.25 packet driver perform the following steps:
trace -a -j 25c
sna -s sna
trcstop
trcrpt > /tmp/pmr/trcfile.fmt
cp /var/adm/ras/trcfile/tmp/pmr