Unknown Error in NDMP subsystem

[O_Loomann]

Hi !

We experience trouble while browsing a Linux (fedore core6) system via ralus from an BE 11d Server. After configuring the user credentials (using root) we get an error (in german language :( )

Similar trnalation: An error occured while browsing <name of the linux host>. Unknown error in the NDMP subsystem.

We need to backup the machine tonight and would really appreciate some help.

thanx in advance,

Oliver

[Lou_Buffetta]

Hi O_Looman

It sounds like you may have a port issue with NDMP. See if the TechNotes listed below help.

The first TechNote (link below) describes your error but only cites Backup Exec 10d. The ports are the same in version 11d so this resolution will apply. Most of the TechNotes refer to RedHat Linux so you will have to take the information and apply it to your particular OS.

How to change the NDMP communication port of the Remote Agent for NetWare with Backup Exec 10.x for Windows Servers

http://seer.entsupport.symantec.com/docs/276745.htm

Unable to start the Backup Exec 10.x Remote Agent for Linux and UNIX Servers (RALUS) on a RedHat Linux system.

http://seer.entsupport.symantec.com/docs/284480.htm

How to determine the Network Data Management Protocol (NDMP) port status on the remote Linux (Redhat) system in Backup Exec 10.x and 11.x for Windows Servers.

http://seer.entsupport.symantec.com/docs/284362.htm

How to change the port for Network Data Management Protocol used by Backup Exec Remote Agent for Linux and UNIX servers on a UNIX/Linux machine

http://seer.entsupport.symantec.com/docs/275228.htm

What TCP/UDP ports does Backup Exec 11d for Windows Servers (including CPS and DLO) and Backup Exec System Recovery use?
http://seer.entsupport.symantec.com/docs/285830.htm

Finally I discovered this recent information

When creating a selection list for backup, attempts to browse a platform running the Remote Agent for Linux/UNIX Servers (RALUS) fails with an error An error was encountered while attempting to browse the contents of.... The media server could not connect to the remote computer. The remote computer may not be running the Backup Exec Remote Agent for Windows/Linux. Make sure that the correct version of the Remote Agent is installed...

Upon investigation, it is found that RALUS is not binding to the NDMP port on the IPv4 stack.  It is doing so on the IPv6 stack.  This may be confirmed using netstat -a.  This occurs no matter what is defined for the NDMP port in /etc/services.  The RALUS debug has the error, "Can't start listener: Error Code: 125, System Error Message: Address already in use".  

It may be possible, however, to utilize the legacy agent for your particular platform to obtain backups.  Please contact Technical Support for assistance with this option.

Here is the link to the Legacy UNIX agent but keep in mind that I have not had any success getting this agent to work with version 11d. It seems 11d only works with the RALUS agent.

VERITAS Backup Exec (tm) 10.0 for Windows Servers - Agent for UNIX version 5.046 (Legacy Agent)

http://seer.entsupport.symantec.com/docs/269803.htm

[O_Loomann]

Hi Lou,

thankx for your suggestions. We have changed the ndmp-Port on both, the Linux-box running the RALUS and the Windows2003 box running BE11d without success. Still the same error message. But one thing we found interesting:

in the old scenario (ndmp on port 10000 on the ralus) we were unable to get an answer by telnetting to this port. After changing the ndmp port on the linux-machine to 12000 we got an answer when telnetting to this port.

As far as i understand it, making changes to the c:\windows\system32\drivers\etc\services on the Windows-Box running BE11d does not affect our topic, because ist only changes the port, beromte on the windows-box is listening to.

What we are wondwring is, how the windows-Box knows on which port to connect to the ralus-machine.

Any idea?

Oliver

[Lou_Buffetta]

When changing port 10000, you need to change each and every server with a remote agent on it. That's because when a job is initiated by Backup Exec, the first communication is sent from the Job Engine to the Remote Agent over port 10000 (or if changed the new port). Once that communications is established, the two services will negotiate the data transfer over a different set of ports.

The Data Connection can change from job to job or even set to set within a job. The Command Connection will always set up on port 10000 going to the remote box.

If another application is already using port 10000, the Remote Agent can't grab it to listen for commands (or advertise itself). The Command Connection cannot be established and Backup Exec will think the Remote Agent is not installed.

[O_Loomann]

Hi Lou,

we understand, that every box with a remote agent needs to be set to the same port, but we would expect, that we also have to tell the Job Engine, which port should be used to initiate the communication with RALUS.

We have only one server with RALUS, where beremote is definitely binding to the selected port (we checked that using the referring Tech-Doc from symantec), so we think that our problem is on the JobEngine side.

We are not sure that changing NDMP Port on the JobEngine-Box (which is Windows2003) to the same different port like in RALUS (e.g. 12000) does really make the Job Engine initiate the connection to the ralus box via Port 12000.

Maybe we also need to tell the JobEngine, which port to use to connect to RALUS ?!?

Oliver

[ChadG]

hello,

Can anyone tell me if this has been fixed yet?  We are doing the same thing (w/same versions) and have the same problem.

thank you

[ChadG]

can someone from support please reply ... thank you.

[mrtubbi]

Hi,

Not sure wether this will help.
I upgraded to 11D on the Media Servers and subsquently found that I needed to upgrade the UNIX agent from the old bexec agent to the RALUS 11D on our Solaris10 Servers.
Tried it out with the same types of results as above....  bummer.
On the media server went to TOOLS/OPTIONS/NETWORK&SECURITY  and checked "Enable remote agent TCP dynamic port range".
I am running backups OK now.
By the way, dont enable "backup contents of softlinks" in your UNIX options for the backup jobs. I found in Solaris 10 this causes the agent to coredump and stop (observer this while running "/opt/VTRSralus/bin/beremote --log-console".
Hope this might help in other *.IX versions

Tim

[Allen_K]

---

ChadG wrote:

can someone from support please reply ... thank you.

---

Hi,

Sorry no one has been able to help you with your technical issue so far. Please note that this is a peer-to-peer discussion forum and not technical support.

Usually customers come to the rescue. Other times, when they are available, a Symantec employee will volunteer his/her time to find and help solve issues just like any other member of the community.

However, in this case, it looks like you might be better off contacting Symantec Support Services for 24/7 access to technical support professionals. You can also try using our Knowledge Base Search.

Hope that helps!