Back story

The network team in my environment is trying to retire an old DNS server. They actually did it assuming nobody was using it at the time. Two of my applications, both running on Weblogic/RHEL, failed, giving several errors about finding other servers needed for running the application, like db and ldap hosts. I couldn't ping these other servers from the weblogic nodes, so we started looking into DNS.

We looked at the /etc/resolv.conf, and only one of the servers in the cluster actually had a valid DNS host in one of the nameserver entries, with the old/retired DNS server still in the list. We removed that server and added newer, valid nameservers, and tried to restart the applications in weblogic, but they did not start up, referencing similar errors as before. The only thing that restored the applications was restoring the DNS server.

Current Issue

We're trying to ensure the applications will run without that specific DNS server, so I'm currently working on the assumption that Weblogic has some DNS IPs cached and causes any resource lookups to go through that list, which only currently contains an actual server because they brought it back up - they still want to retire it though... how do I get Weblogic to definitely and certainly use the newer list of DNS nameservers that I've got (which I've added to the resolv.conf)

Is there a place to update a list in Weblogic? should it be getting the nameservers from the host OS? do I have to restart Weblogic? Node manager? both?

  • In general the work-around for dns issues is adding the the records required for your application to function correctly to your hosts file, overriding DNS name resolution completely. How viable that is depends on the situation.
    – Bob
    Feb 24 at 21:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.