The following new enhancements have been added in this release:
Enhanced reporting of Service Time-Out errors provides the following new capabilities:
Reporting of Service Time-Out Errors
New Capabilities
The Service Time-Out enhancement includes more refined error detail (TPED_SVCTIMEOUT) and a new system event (.SysServiceTimeout). The client now receives additional information indicating that a service time-out has occurred when BEA TUXEDO terminates a server process due to a service time-out (configuration parameter SVCTIMEOUT in the UBBCONFIG file, TA_SVCTIMEOUT of T_SERVER and T_SERVICE in the MIB).
Formerly, only the error code TPESVCERR was returned when a service time-out occurred. Because the SVCTIMEOUT value is configurable, it is important for the client applications to easily distinguish a TPESVCERR that may be caused by the value set for SVCTIMEOUT, from those caused by other situations resulting in TPESVCERR. Although the ULOG contained this information, it was difficult for the client program to extract it. To differentiate the service time-out TPESVCERR from others, a call to the new tperrordetail () routine after detecting a TPESVCERR, yields TPED_SVCTIMEOUT when a service time-out occurs.
Also a new system event, .SysServiceTimeout is generated when a service time-out occurs. If the .SysServiceTimeout occurred, it is reflected in the ULOG in the following way: Sometimes an unexpected system error occurs that freezes a service or causes it to run out of control while it is processing a request. Though desirable to remove these processes, it is difficult to detect them or their origin. A BEA TUXEDO mechanism terminates these processes based on the time it takes for a service to process a request.
You can configure the length of time (service time-out) through the SVCTIMEOUT parameter in the TUXEDO UBBCONFIG file or by dynamically changing the TA_SVCTIMEOUT attribute in the MIB. By default, BEA TUXEDO does not terminate any service process, so you must set the SVCTIMEOUT value (in seconds) to activate this feature. We recommend that you set the value of SVCTIMEOUT or TA_SVCTIMEOUT to at least two to three times greater than the time it takes for the longest running service to process a request. Setting the service time-out this way guarantees that BEA TUXEDO removes only frozen processes. In essence, the service time-out acts like a scavenger for frozen or out of control application servers.
When a time-out occurs, BEA TUXEDO terminates the server process running from the frozen service (but not the child processes). It then returns the TPESVCERR error indicating that an unknown problem occurred during the processing. In a conversational service, the conversation event TPEV_SVCERR is returned.
TPESVCERR or TPEV_SVCERR is returned in the following situations:
More Refined Error Information
ERROR: .SysServiceTimeout: %TA_SERVERNAME, group %TA_SRVGRP, id %TA_SRVID server killed due to a service time-out
How BEA TUXEDO Reports These System Errors
What Happens When a Time-Out Occurs
Situations that Cause a Service Time-Out
How to Control a Service Time-Out
Although no default device for TLI machines existed prior to this release, as of Release 6.4, you no longer need to specify the device for TLI machines. If you do not want to use the default, you can still specify a device which is platform-dependent (for example, dev/tcp).
Specifying the -d <device> is no longer required to start processes such as tlisten, wlisten, WSL (CLOPT), tuxwsvr, etc:
Default Device for TLI Added
When running multiple BEA TUXEDO domains simultaneously on the same machine, it was sometimes difficult to match BEA TUXEDO processes with their associated domains using the ps command or winps on Windows NT.
To enable quick and easy association between BEA TUXEDO processes and their respective domains, Release 6.4 provides a new DOMAIN ID to all processes. When a DOMAINID is specified, a comment option -C dom=<domainid> is present on the command line for all servers.
Domain ID on the Server Command Line