I have an api gateway cluster that has 2 gateways. The gateways talk to an ldap server.On both gateways, networkaddress.ttl=-1 is commented out in java.security file.
When ldap's IP address is changed in DNS, one gateway will start to use the new IP but the other gateway continues to use the old IP (and therefore can't connect to ldap).
Has anyone seen similar issues? Any solutions?
The default networkaddress.ttl is forever. But when a security manager is not set, the default is 30 seconds. How do I know if I have security manager set or not set?
Does each gateway handle this individually or it's handled by the cluster and the networkaddress.ttl setting is replicated between the 2 gateways?
To add some info, my gateway is 9.3 in AWS environment. There's ELB in front of my ldap nodes. The DNS service on the gateway nodes resolves ldap hostname to the ELB for the ldap.
Is this a bug in the api-gateway software?
As you already know, the DNS caching problem is from the java platform and it's expected.
For AWS servers, we recommend to use Elastic IP for them, you may refer to Amazon document,
Thanks for the reply.