Long system-wide pauses when CIFS server unreachable

Long system-wide pauses when CIFS server unreachable

[R. Diez]

Hi all:

Today I forgot again to unmount a CIFS mount before turning my Windows 10 PC off. After resuming my Linux laptop from sleep, the file manager and other applications froze for a number of minutes.

Unmounting the unresponsive mount was difficult. Command "lsof | grep mountpoint" showed nothing. I am using echo_interval=4, but that does not help.

I reported this issue here some months ago, and got some feedback from Lucy Kueny, but otherwise CIFS maintainers were silent:

Subject: How to automatically drop unresponsive CIFS /SMB connections
Date: Sat, 3 Feb 2024 23:48:42 +0100	
https://lore.kernel.org/linux-cifs/428ab7ba-0960-4e5e-a4ab-290dac58f45b@rd10.de/

I have been experiencing this problem for years, both at home and at work, with a variety of Linux and Windows versions. It is a well-known pitfall.

I think that this is actually a serious issue. Most users wouldn't know why the Linux system is suddenly unresponsive, or how to fix it. Even restarting a system with a hanging CIFS connection is problematic.

I am using Ubuntu 22.04.5, kernel version 6.8.0-45-generic.

Is there some new work-around available nowadays? Otherwise, is there any chance that the CIFS maintainers will take this problem seriously?

I am considering switching to OpenSSH for Windows, but I can only do that with my personal Windows computers. NFS for Windows is probably not an option. In a business environment, most Windows sysadmins will not install other network protocols anyway just because Linux has issues, so there is not much most Linux users can do.

Regards,
   rdiez