This content has been marked as final. Show 1 reply
You are describing 2 different problems althoguh they may be related, but I'll answer them in separate context.
The first problem:
You listed 2 related Tuxedo error messages. What these 2 error messages say that GWTDOMAIN gateway failed to forward reply to Tuxedo client. The most common situation to have those 2 error messages logged in ULOG was because client is not available, or in another words client process already terminated. This client termination may be because service time out or transaction time out occurred so the client received failure reply from local BBL process, this can also caused by client was killed or tmshutdown was performed.
The second problem:
The most likely cause of this mismatched buffer is because configuration mismatch. One question you need to ask yourself is whether EJB received a FML32 buffer with recognizable field id, i.e. the field id defined in other FML32 or FML buffer field table. To determine if this is the case you can compare the field id in the header files generated from all your FML/FML32 field tables. If you can find them in the field table header files then it quite possible you have service name configuration error.
In /Domain DMCONFIG file you configure DM_IMPORTS (or DM_REMOTE_SERVICES) for the resource resides in WebLogic Server, each entry in DM_IMPORTS must have remote resource name optionally can have "RNAME". When GWTDOMAIN gateway forward the request to WTC it will check whether RNAME is configured or not, if configured then it will use RNAME as target resource name when forwarding the request, otherwise it uses the configured remote resource name when forwarding the request to WTC.
The similar thing happens in WTC. In WTC you configure Export for exported resource; and optionally, you can configure RemoteName. So a reverse translation may occur.
Lastly, check the JNDI name that bound your EJB to is correctly configured and matching the information in weblogic-ejb-jar.xml.