Reload (all versions)
The customer's post office server went down so the disaster recovery (DR) POA on the Reload server was launched. Users could connect to the DR POA just fine and were sending and receiving messages; however, the administrator noticed the production MTA was showing that the post office was closed and that it was trying to connect to the production POA's IP address, which was down. He sent a test message from a live PO to the DR post office and it was never received.
- Checked the POA's network address in ConsoleOne and it was set with a DNS name, which was correct.
- Rebuilt the domain database. Same issue.
- Tried to ping the POA using the DNS name: it returned the correct POA address on the Reload server.
- Tried to telnet to the POA using the DNS name and the port the POA was listening on (1677): the connection was refused and telnet returned the IP address of the production POA.
The DNS cache was corrupted on the MTA's server. We flushed the cache by stopping the name server cache daemon ("rcnscd stop") and starting it back up again ("rcnscd start"). We then restarted the MTA and it was able to connect to the DR POA. The test message sent from the live post office was then delivered to the user mailbox on the DR post office.
This article was originally published in the GWAVA knowledgebase as article ID 2144.