F17: NFS woes

chaud lapin · #1 22nd July 2012, 01:00 PM

Hi all,

There seem to be various posts about NFS problems under Fedora 17. I've just browsed them and not found the solution to my situation.

I have been using NFS shares for many years under Fedora. After I upgraded my desktop box from F14 to F17, my client running Mint 13 can no longer access the NFS shares. I am not using a firewall on the server.

Just now, I have followed to the letter the instructions given here to no avail: http://www.server-world.info/en/note?os=Fedora_17&p=nfs

My export file looks like this:

Code:

/media/disk1        192.168.2.0/24(rw,sync,no_root_squash,no_all_squash) 
/media/disk2        192.168.2.0/24(rw,sync,no_root_squash,no_all_squash) 
/media/disk3        192.168.2.0/24(rw,sync,no_root_squash,no_all_squash) 
/media/disk4        192.168.2.0/24(rw,sync,no_root_squash,no_all_squash) 
/media/disk5        192.168.2.0/24(rw,sync,no_root_squash,no_all_squash) 
/home/colin         192.168.2.0/24(rw,sync,no_root_squash,no_all_squash)

When I try to mount a share on the client, all I see is this:

Code:

mount.nfs: timeout set for Sun Jul 22 13:24:29 2012
mount.nfs: trying text-based options 'vers=4,addr=192.168.2.2,clientaddr=192.168.2.4'

Nothing further happens.

I have tried mounting a share locally on the server; this doesn't work either:

Code:

sudo mount -v -t nfs localhost:/media/disk5 test-nfs
mount.nfs: timeout set for Sun Jul 22 13:48:28 2012
mount.nfs: trying text-based options 'vers=4,addr=::1,clientaddr=::1'

Any ideas what is going wrong here?

#2 22nd July 2012, 01:08 PM

Post

Code:

systemctl status nfs-server.service

and

Code:

iptables -L

jpollard · #3 22nd July 2012, 01:36 PM

There have been a number of changes to the system.

1) /media is a tmpfs now. This means that any mountpoints there get lost on every reboot.

2) NFS depends on the network. Unfortunately, systemd + networkmanager don't by default allow NFS. systemd assumes the network is available as soon as networkmanager is started... and it takes networkmanager several seconds before it may be ready.

3) systemd doesn't synchronize things anymore. Everything is attempted to be processed at the same time.

#1 should cause mount failures (if it is the disk server).

The effect of #2 and #3 are that you have to enable a dummy service (NetworkManager-wait-online.service) to cause systemd to wait for NetworkManager to tell it that the network is ready. Usually this is client related but having the network unavailable may cause issues on the server when the exports are loaded (the network won't be available during the initialization).

The problem with #1 is that unless you recreate the mount points on every boot it will not be mounted.

For servers, you need to check that SELinux has NFS exports enabled, and that the firewall has NFS enabled on the server.

chaud lapin · #4 22nd July 2012, 03:47 PM

Code:

# systemctl status nfs-server.service
nfs-server.service - NFS Server
	  Loaded: loaded (/usr/lib/systemd/system/nfs-server.service; disabled)
	  Active: active (exited) since Sun, 22 Jul 2012 13:34:09 +0200; 3h 9min ago
	 Process: 21095 ExecStopPost=/usr/sbin/exportfs -f (code=exited, status=0/SUCCESS)
	 Process: 21092 ExecStop=/usr/sbin/rpc.nfsd 0 (code=exited, status=0/SUCCESS)
	 Process: 21116 ExecStartPost=/usr/lib/nfs-utils/scripts/nfs-server.postconfig (code=exited, status=0/SUCCESS)
	 Process: 21104 ExecStart=/usr/sbin/rpc.nfsd $RPCNFSDARGS ${RPCNFSDCOUNT} (code=exited, status=0/SUCCESS)
	 Process: 21101 ExecStartPre=/usr/sbin/exportfs -r (code=exited, status=0/SUCCESS)
	 Process: 21098 ExecStartPre=/usr/lib/nfs-utils/scripts/nfs-server.preconfig (code=exited, status=0/SUCCESS)
	  CGroup: name=systemd:/system/nfs-server.service

Code:

# iptables -L
Chain INPUT (policy ACCEPT)
target     prot opt source               destination         

Chain FORWARD (policy ACCEPT)
target     prot opt source               destination         

Chain OUTPUT (policy ACCEPT)
target     prot opt source               destination

Quote:

/media is a tmpfs now. This means that any mountpoints there get lost on every reboot.

Not following you entirely here, jpollard. I've added these /media mounts to fstab (on the server) and they come up when the system is rebooted.

#5 22nd July 2012, 04:01 PM

Try ( any mount points in /media are destroyed on reboot as /media is meant for temporary mounts only), use /mnt instead!!!

Code:

su
mkdir /mnt/disk{1,2,3,4,5}
sed -i -e 's,media,mnt,g' /etc/fstab /etc/exports
systemctl enable nfs-server.service
reboot

chaud lapin · #6 22nd July 2012, 05:06 PM

Thanks Yellowman. Moving my mounted drives to /mnt and exporting from there solved everything.

One small detail: on the desktop box (the NFS server), the mounted drives are no longer appearing on my Gnome3 (actually Cinnamon) desktop. There is a setting under Settings/Desktop: "Show mounted volumes on the desktop" which is (still) selected, but I don' t see the drive icons. Any thoughts?