0

I'm getting this error every 5 minutes on one of my Linux servers. The name in this case is not a hostname, but the workgroup-name I have set in the smb.conf file of both servers. I have no idea why it keeps doing this, the only thought I have (as a Windows admin) is that the second server is trying to become the master-browser or something and upon resolving the workgroup name gets a reply front the first server that was turned on. Which would explain why the other server shows no errors in its log.

Both systems have statically configured IP's and resolv.conf containing both domain and nameserver entries.

It makes me wonder though: since it is nmbd that's doing this, do I still really need it in a SMB3.1 Win10 network? Or can we never get rid of NETBIOS?

1

It makes me wonder though: since it is nmbd that's doing this, do I still really need it in a SMB3.1 Win10 network? Or can we never get rid of NETBIOS?

nmbd's functions (browsing and NBNS) have always been optional for modern clients (Win2000+). The last system which deliberately sent a NetBIOS datagram query before the actual SMB connection was Windows 98 (or perhaps Windows ME).

So you can connect to the smbd service by IP address (or DNS name) and speak whatever SMB/CIFS version you like, whether it's SMBv1 over NetBIOS Session over port 139, or whether it's SMBv3.1 over raw TCP over port 445, without having nmbd running.

  • As a primarily Windows admin, I'm often confronted with the fact that even with Windows 10 I must still open up the NETBIOS ports to access file shares. I guess if I only access my smb-shares using a FQDN, it should not have to use NB. The fewer services I need running, the better anyways. I'll try and disable nmbd for a while and see if I run into weird problems.. – Mark May 26 at 10:52
  • As long as you have any alternative method of resolving the name (raw IP address, /etc/hosts, DNS, mDNS, LLMNR) you should be fine with only having smbd and only port 445/tcp. (SMBv2+ isn't supposed to even use 139/tcp.) – grawity May 26 at 13:23
0

Multiple responses could be received because of a host in the same physical network which has two interfaces. The broadcast query is then replied to both of them by your nmbd. See this 17 years old post: query_name_response: Multiple responses received.

Check also the file smb.conf regarding the interfaces parameter, if it doesn't list more than one interface.

  • I already checked, but systems have only a single interface. For that reason, I have not set the interface-property in smb.conf. The confusing part is that the name that is referred to in the error is the name of the workgroup and not either system's hostname. – Mark May 26 at 10:50

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.