I have a RHEL system that's been connecting to a Windows file share through NFS, but I'm facing some strange issues. This setup is pretty old, coming from a Solaris system over a decade ago. The file shares are organized by department for end users, and the accounts on Windows match the RHEL/Unix UIDs for security purposes. Lately, the NFS mount on RHEL has been randomly stalling, and we can't force a remount. Restarting the Windows service doesn't work either; only a complete reboot of the file server helps. Interestingly, all Windows clients can access the share without any problems during these stalls. The logs aren't giving much insight either. I'm considering switching to SMB instead, although I understand that it'll add encryption and require connections to default to the user specified in the mount command instead of the RHEL user. Does anyone foresee any issues with this switch?
4 Answers
Have you checked for any conflicting IP addresses on the network? Sometimes that can cause connectivity issues, though it sounds like yours may be specific to the NFS connection.
You might want to run a rotating tcpdump on the Windows file server to see what last operation is happening just before it stops responding. That could give you some clues. I've used the NFS Server on Windows before and faced issues, so switching to CIFS might be a better option for you. Just keep in mind that you'll need to manage credential rotation if you do switch.
It sounds like it might be time to upgrade your old setup. What's your current Windows infrastructure like? If you're on an Active Directory domain, integrating your RHEL system could streamline things so your users only need one set of credentials without needing local accounts. Generally, SMB is simpler for file sharing in Windows. Encryption isn't required unless you set it up for SMB3. If you're on a Red Hat desktop, maybe just create a shortcut to the share in the file explorer for easier access.
We do a lot of NFS from Windows Server, and I haven't encountered issues like yours. When the stalls happen, try running `rpcinfo -p ` and `showmount -e ` on the Linux clients to see if those commands return results quickly like they do when the server is freshly booted. Consider these questions:
* Do all clients stall together?
* Are you using any connection-tracking firewalls?
* Is there real-time file-scanning antivirus on the Windows side?
Yep, that's not the issue. It definitely seems tied to the NFS setup.