The hard drive on my home Linux server crashed on me today. But I’ve been backing it up to myCloud every day via rsync to an NFS share. Boy was I surprised to discover that it actually hasn’t been backing up since January 20 of 2019. Over a year ago. So what happend? First I discover I can no longer log in as root via ssh, it’s now sshd. And even thought I have no_root_squash and no_user_squash set up in the /etc/exports, they don’t actually seem to have any affect. So obviously this quietly started happening Jan 20, 2019, but now how do I turn it back on. A bit upset that I’ve lost a year’s worth of mail and modifications to my web server, but I can cope. I wold just like to have my back up device working again the way I bought it. (current FW:
2.31.204)
Do you really want root@Linux_server = root@myCloud and user@Linux_server = user@myCloud? That is what the no_squash parameters do - you have to have the same users on both machines.
yes, that is really what I want to do. I am rsyncing to it via an NFS share. If the root user gets squashed to nobody then I can’t read or write my system files to the backup NAS and maintain ownership.
The problem is that the nfsd doesn’t seem to be obeying my squash settings
Are you sure it was the nfsd? myCloud apparently removed p/w for root in FW upgrade (I am on 4.x FW), so that may have broke rsync. Did you use the “user@host” option for rsync?
Not even trying to use an rsync server on the myCloud. Just plain NFS. If I mount the share, via NFS, I cannot read or write to it anymore. I always get permission denied.
I’m currently restoring files via a tar pipe through and ssh connection from the myCloud to the new server. It works fine for recovery, but I still need to be able to write to the NAS, maintaining permissions, using a native *nix mode.
And I really don’t want to use tar. It’s a disk, not a tape. I really don’t want to create an incremental serial backup of my hard drive. That’s so 1980’s.
futher digging, I discovered that the reason no_[all_|root_]squash does not work is because the myCloud forcibly rewrites /etc/exports when the daemon is restarted. And you can’t disable that because the nfs start script that does it is in rom or something. It’s a read only fs called loop0.
If you meant can not disable. If you are on a gen2 which I think you are. Search this forum for funplug.
It is available on the gen2. It allows you to run a script after the system boots up.
I’ll give it a shot, but when I tried to “exportfs -r” after the daemon was running, the client claimed that the nfs version was unsupported. More keyboard banking later.
