Datacom Server client using ODBC connection receives a sporadic error and with no changes to the application. The error from the application is SQLSTATE 08001, but the driver has error TCP(11002) WSATRY_AGAIN with Nonauthoritative host not found.
Microsoft Windows client
Here is a simplified explanation of that error, as it relates to the TCPIP code 11002:
Nonauthoritative host not found: This error suggests that while the DNS server queried can access some data about the domain, it lacks authoritative details from the primary DNS server used for lookup. This issue might stem from temporary connectivity problems with the authoritative DNS server. It is important to discuss with your networking team to review the DNS configuration of your server.
As for the subsequent message "WSATRY_AGAIN", you can search this and find:
This is usually a temporary error during host name resolution and means that the local server did not receive a response from an authoritative server. A retry at some time later may be successful.
This error seems to be unrelated to your application, and appears to be taking place in the DNS processing layers, where the resolver could not contact the name server or got some kind of bad response from the name server. Error 11002 is a common error encountered when an application or service cannot contact a DNS server to resolve a domain name, or when the DNS server is misconfigured, or simply does not have the record for the domain in question (this last item is not likely, as your error is sporadic).
If one machine is more prone to these errors than others, or when you run into this error (assuming your client is running Windows), you should verify your network settings or contact your network team for assistance. One thing to check is that you do not have an incorrect entry for the server in your hosts file (use Command Prompt command NOTEPAD %SYSTEMROOT%/System32/drivers/etc/hosts). If there’s an incorrect entry, it could cause a DNS resolution failure, but as noted above, since this error is sporadic, this is not likely to be the problem.
If you want to flush your current DNS cache and rebuild it, you can issue the Command Prompt command ipconfig /flushdns which will delete the locally stored DNS records, forcing the system to re-query the DNS server for fresh records. Use ipconfig /displaydns first if you want to see what entries you have in your cache.
This does not appear to be related to your Server application (on mainframe or client), but seems to be a network issue.
As always, please contact Broadcom support for Datacom if you have further questions.