ORA-19511 / NetBackup Status 25: Cannot Connect on Socket

ORA-19511 is Oracle RMAN reporting an error returned by the media-management layer. When the accompanying NetBackup message is status 25: cannot connect on socket, troubleshoot the NetBackup connection rather than the Oracle database first.

What status 25 means

A NetBackup process timed out while connecting to another NetBackup process. Common causes include name-resolution errors, a required daemon or Windows service that is not running, blocked connectivity, heavy server load, or an expired license. The Oracle error is the wrapper; the NetBackup status and logs identify the underlying problem.

Troubleshooting checklist

1. Capture the exact time and hosts involved. Note the RMAN channel, database client, media server, and primary server named in the job details.
2. Verify name resolution in both directions. Each host should resolve the other to the expected address. Check DNS and local host files for stale, short-name, or duplicate entries.
3. Verify NetBackup services. Confirm the request, database-manager, client, network, and PBX components required by your NetBackup release are running. On UNIX or Linux, bpps -x is useful; on Windows, check the NetBackup services.
4. Test the client connection. From the appropriate NetBackup server, run bptestbpcd -client <client_name> -verbose. Use bpclntcmd -hn, bpclntcmd -ip, and bpclntcmd -self to isolate host-name problems.
5. Check firewalls and routing. Confirm that the ports required by your installed NetBackup version are allowed between the specific client, media server, and primary server.
6. Read the matching logs. Review the logs for the processes named in the failure, such as bprd, bpdbm, bpcd, vnetd, bpbrm, or PBX. Start with the first connection error at the same timestamp as the RMAN failure.
7. Check load and licensing. A slow or saturated server can make a healthy connection time out. Also verify that the NetBackup license is valid.

About TIME_WAIT tuning

Older troubleshooting notes often recommend reducing the operating system’s TCP TIME_WAIT interval. Do not make that the first fix. Change kernel or registry networking values only when logs and socket statistics demonstrate port exhaustion, and only after reviewing the setting with the operating-system and network teams. An incorrect global TCP change can affect every application on the host.

Reference

See the official Veritas NetBackup status code 25 reference. Commands and required services can vary by NetBackup release, so pair this checklist with the documentation for the installed version.