NFS automount /etc/fstab with Jessie

[wopper]

Hi!

Im installing a fresh new Jessie image to my second Pi, a linux beginner im copy the config from the running Wheezy on to the new Jessie. So far so good but whith the NFS mounts in /etc/fstab it will not start, an mount by cli works like a charme.


Old and working wheezy /etc/fstab config:

Code: Select all

172.16.1.2:/volume1/pi /mnt/backup nfs rw,relatime,vers=3,rsize=131072,wsize=131072,namlen=255,hard,proto=tcp,timeo=600,retrans=2,sec=sys,mountaddr=172.16.1.2,mountvers=3,mountport=892,mountproto=udp,local_lock=none,addr=172.16.1.2

Exports on the synology:

Code: Select all

/volume1/pi	172.16.1.0/24(rw,async,no_wdelay,no_root_squash,insecure_locks,sec=sys,anonuid=1025,anongid=100)

if i mount by cli it works with this command

Code: Select all

sudo mount -t nfs  172.16.1.2:/volume1/pi /mnt/backup

Any idea why its not working, and where to find logging for it? tnx

[wopper]

Changed the nic settings from DHCP to static and it works fine, looks like a timing issue when loading the different services.

[PeterO]

Changed the nic settings from DHCP to static and it works fine, looks like a timing issue when loading the different services.

But that is not a generally applicable solution Systemd should be able to deal with this but doesn't
PeterO

[dom]

I hit this issue and with a lot of searching found adding ",noauto,x-systemd.automount" to the end of the options fixed it for me.

[DougieLawson]

I hit this issue and with a lot of searching found adding ",noauto,x-systemd.automount" to the end of the options fixed it for me.

That's the solution I'm going to test this weekend when my main 24/7 running server gets a reboot for Jessie. I've got two systems already running Jessie (although one has crashed and I'm 80 miles from home so it hasn't been restarted). Four other systems are still running Wheezy and will be upgraded this weekend if that test works.

[Cecil]

Yes! the noauto,x-systemd.automount options do work!

Thank you

[dom]

An alternative that seems to work too is to launch dhcpcd in a systemd friendly way.

Code: Select all

sudo rm /etc/init.d/dhcpcd
cat | sudo tee /etc/systemd/system/multi-user.target.wants/dhcpcd.service << EOF
[Unit]
Description=dhcpcd on all interfaces
Wants=network.target
Before=network.target

[Service]
Type=forking
PIDFile=/run/dhcpcd.pid
ExecStart=/sbin/dhcpcd -q -t 0
ExecStop=/sbin/dhcpcd -x

[Install]
WantedBy=multi-user.target
EOF

Now it seems to mount without ",noauto,x-systemd.automount".
Note: assuming this is reported as working, there will be an update to make this the default.

[Dandelion]

The use of the "noauto,x-systemd.automount" instructions in fstab works here too. Looks like the dependencies at boot are not correct (rpcbind not starting until after the nfs mount failure). Once booted, the system will mount the NFS share no problem using:
sudo mount -a
Looks like I need to do a bit of reading about how to influence boot sequencing and how to add explicit dependencies under systemd.

[timrowledge]

Just as an extra datapoint, adding the neat stuff works nicely for me too. I end up with -

Code: Select all

192.168.1.65:/Users/tim /home/pi/DizietFS nfs rsize=8192,wsize=8192,timeo=14,intr,noauto,x-systemd.automount 0 0

[rbn]

This is great. I've been trying to get auto mount of my nfs share to work since moving to Jesse, and it always hung because the wireless network was not fully up when it was called.
adding noauto,x-systemd.automount to the fstab entry sorted it.
Many thanks

[Malakesher]

An alternative that seems to work too is to launch dhcpcd in a systemd friendly way.

Code: Select all

sudo rm /etc/init.d/dhcpcd
cat | sudo tee /etc/systemd/system/multi-user.target.wants/dhcpcd.service << EOF
[Unit]
Description=dhcpcd on all interfaces
Wants=network.target
Before=network.target

[Service]
Type=forking
PIDFile=/run/dhcpcd.pid
ExecStart=/sbin/dhcpcd -q -t 0
ExecStop=/sbin/dhcpcd -x

[Install]
WantedBy=multi-user.target
EOF

Now it seems to mount without ",noauto,x-systemd.automount".
Note: assuming this is reported as working, there will be an update to make this the default.

I would just like to say that after over 3 hours of trying to successfully mount a networked drive (which should have taken me a couple of minutes) your solution is the only one that worked for me.

Including ",noauto,x-systemd.automount" not working.

Thank you very much and I hope this gets fixed very soon.

[rubikwizard]

Thanks Dom

Code: Select all

,noauto,x-systemd.automount

This worked for me too. Only 'problem' is that the mounted volume appears twice on my desktop (using different icons). It only used to appear once. Both of them work. Any idea how to remove one of them?

[wopper]

Hi guys

i reinstalled the Pi again, as a newbee you make mistakes

But i'm unable to autostart "rpcbind" with RC.D because of the missing RC config parameters in the default rpcbind script, i now inserted it to rc.local and that works.

But i can't stand it, that i had it working on Jessie with the previous installation, could you tell me how to start the NFS service in a decent way?

The " noauto,x-systemd.automount " works fine with reboot.

[DougieLawson]

You shouldn't need to do that. Adding rpcbind to rc.local is a bad idea.

If you've got all of the NFS stuff installed correctly then

Code: Select all

192.168.1.230:/nfs/home /home  nfs defaults,noatime,noauto,x-systemd.automount 0 0
192.168.1.230:/nfs/shared /shared nfs defaults,noatime,noauto,x-systemd.automount 0 0
192.168.1.230:/nfs/server /srv nfs defaults,noatime,noauto,x-systemd.automount 0 0

should encourage your Jessie systemd to start rpcbind for you. The important thing is that x-systemd.automount clause.

With Wheezy you'd enable rpcbind with
sudo update-rc.d rpcbind defaults
sudo update-rc.d rpcbind enable

Those commands should work with your Jessie systemd installation to create the right systemd service files.

[wopper]

Hi Dougie

you completely right! It works even without the rpc bind rc.local or rc.update code.

Many thanks! (it even works with DHCP)

[nitrosont]

I've got the similar problem. At my Raspberry running Jessie I cannot add a nfs share.
I put the following in my /etc/fstab

Code: Select all

192.168.1.3:/volume1/folder /home/pi/nas/folder nfs defaults,noatime,noauto,x-systemd.automount 0 0

This results in the following: I get the folder "folder" with owner:group = root:root.
Of course I cannot enter the folder and not access it.

Anyone an idea?

[amb1ent]

I can confirm only the systemd configuration file change worked for me, the fstab options had no effect.

[sancho_mic]

hi,

Got exact same issues as OP.

I tried the option from Dougie - no luck...But it worked on wheezy. Now on jessie not working anymore...
EDIT: had to reverse "wheezy" with "jessie" in the above statement - got confused which one is older)

I tried with the noauto,x-systemd.automount, no luck...

I tried with the option from dom - no luck

I tried changing the /etc/exports of the nfs server (synology) - the "insecure_locks" vs "insecure" - i'm not even sure if this trick is required on DSM5.2 and for wheezy. I know for sure it was required for openelec to access nfs shares @ synology in the past.


At certain point, my raspberry was behaving in a very quantum-ish way
the systemctl status mnt-mynfsshare.mount command was showing it as failed.

However when i was trying to access the mount it was working. Reissuing the systemctl status mnt check was showing now as active...
So i guess it has to do with the automount thing.2

In my scenario, i run a script @reboot via crontab - this script needs to access the nfs mount.
Script fails and stops because nfs mount fails initially during reboot.

Any suggestion?

Maybe there is better practice than to use the @reboot in crontab, but it was simple to implement - i'm just noob...

Many thanks,
-s

[DirkS]

I tried the option from Dougie - no luck...But it worked on jessie. Now on wheezy not working anymore...

Did you downgrade from Jessie to Wheezy?

[sancho_mic]

thanks for the fast reply DirkS...
i edited my post, got confused with the jessie wheezy thing...

So i'm running jessie, the latest, and i'm facing issues.
Was working ok in wheezy...

[eponymous]

Code: Select all

,noauto,x-systemd.automount

was not working for me and the raspbian jessie image I downloaded was already starting dhcpcd with a systemd service file.

I finally got rpcbind to reliably start by disabling it through systemd, moving /etc/init.d/rpcbind to different directory (it's not clear to me why, but it would otherwise still try to start even after I'd disabled it), and finally by creating and enabling an rpcbind.service file based on https://lists.debian.org/debian-devel/2 ... 00618.html:

Code: Select all

[Unit]
Description=RPC bind portmap service
After=systemd-tmpfiles-setup.service
Wants=remote-fs-pre.target
Before=remote-fs-pre.target
DefaultDependencies=no

[Service]
ExecStart=/sbin/rpcbind -f -w
KillMode=process
Restart=on-failure

[Install]
WantedBy=sysinit.target
Alias=portmap

I also created the directory /run/rpcbind/, which rpcbind seemed to want based on some non-fatal errors it was issuing.

[DougieLawson]

Try this
sudo apt-get purge rpcbind
sudo apt-get install rpcbind nfs-common

[sancho_mic]

Dougie hi

Your purge did the trick!
Now all ok
Mount is done during boot (and my script starts writing to the nfs mount since boot)

Many thanks!
-s

[sancho_mic]

@eponymous
thanks for the suggestion, will give it a try on next reinstall.
cheers
-s

[lsrmgr]

I hit this issue and with a lot of searching found adding ",noauto,x-systemd.automount" to the end of the options fixed it for me.

Dom, thank you for posting this. I works nicely for me with Debian Jessie. I have been trying to get nfs automount to work since installing Jessie when it first came out.