I had no problems until today with sudo in an AppVM. After installing stuff and rebooting, it has changed permissions. Sudo is still owned by root, but permissions are wrong. Sudo is —s–x–x.
Sudo works in my other AppVMs, it is just this one.
Is there a way I can force it to the normal permissions without having to lose that AppVM?
Also, would this —s be why my backup hangs and won’t complete?
You can qvm-run -u root <your-appvm-template-name> xterm and there chmod it, stop template and restart appvm.
I don’t think backup depends on content of the VM. You can also clone the VM before messing with it.
That worked, but I still can’t get sudo working in my AppVM. Can you help further?
When I try to run sudo (after chmod 4755) I get:
sudo: effective uid is not 0, is /usr/bin/sudo on a file system with the 'nosuid' option set or an NFS file system without root privileges?
I don’t know. There must be some advise for the error on internet somewhere. 
Yeah. Now I think it is a Qubes specific issue I’m having. Following the steps you provided, I can alter the sudo owner/permissions in the template, and it changed in my other AppVMs. This particular AppVM still won’t work even though I restarted the templateVM and AppVM.
So it seems that my problem is not from the cause I thought (I thought it was the ownership/permissions of sudo).
Seems that other Linux users had to remount, but I will just hope that a reboot fixes it for me, or I guess I will recreate this AppVM.
mount -o remount,rw /
mount --all