Mounting file system: Failed. Uploading my first image

[Miguel Palacios]

Hi guys,
I am trying upload my first image to my FOG server but… I get this error.

My FOG server is installed in a CentOS 6.5, the base version is 1.2.0 and SVN version is 7009.
So are my services:

[root@IMG-28V-LMES-20 dev]# service iptables status
iptables: El cortafuegos no está ejecutándose.
[root@IMG-28V-LMES-20 dev]# service nfs status
rpc.svcgssd está parado
Se está ejecutando rpc.mountd (pid 14293)…
Se está ejecutando nfsd (pid 14308 14307 14306 14305 14304 14303 14302 14301)…
Se está ejecutando rpc.rquotad (pid 14289)…
[root@IMG-28V-LMES-20 dev]# service rpcbind status
Se está ejecutando rpcbind (pid 14174)…

My internal storage hard disk is mounted in /images:

UUID=“37627050-0f61-4233-8e4f-bba52063d79a” /images ext4 defaults 0 0

This is the file exports:

*/images *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)
/images/dev (rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)

And these are the permissions in the images folder:

— drwxrwxrwx 5 fog root 4096 mar 31 12:04 images
------ drwxrwxrwx. 2 fog root 4096 mar 30 16:38 dev
----------- -rwxrwxrwx. 1 fog root 0 mar 30 16:38 .mntcheck
------ drwxrwxrwx 2 fog root 16384 mar 30 14:21 lost+found
------ -rwxrwxrwx. 1 fog root 0 mar 30 16:37 .mntcheck
------ drwxrwxrwx 2 root root 4096 mar 31 12:04 postdownloadscripts

And my configuration storage in the FOG web is this…

I had tried follow the instructions for Testing NFS according this manual (https://wiki.fogproject.org/wiki/index.php?title=Troubleshoot_NFS) but the result is the same…

I’m lost…

Thank you guys!

[george1421]

from your fog server will you show us the output from this command showmount -e 127.0.0.1?
This will show us what nfs actually exported.

Is this storage node ON the FOG server or is it setup someplace else. I see that you have it set to master, I just want to ensure that it is installed on the fog server.

Are you using the location plugin or is the FOG servers and client computers in the same subnet.

Last question (you were very informative in your original post), did you remember disable selinux?

[Miguel Palacios]

Hi George,
Thank you for your response.

If I execute the command showmount -e 127.0.0.1, this is the result:

[root@IMG-28V-LMES-20 ~]# showmount -e 127.0.0.1
Export list for 127.0.0.1:
/images/dev *
/images *

It’s looking good but… After when I try to connect, via FOG or linux shell doesn’t work…

The FOG server has two hard disk, one hard disk for SO (30 GB, ext4) and other hard disk for images (4 TB, ext4).

The client and FOG server are in the same subnet (10.6.82.X)

By last, the disable configuration is the next:

SELINUX=disabled
SELINUXTYPE=targeted

On the other hand, reading again my first post… I had saw that the rpc.svcgssd is stopped… Is this correct or can this a problem?

Thank you!

[george1421]

@Miguel-Palacios It might be interesting to know if you could do this.

Setup a new image reference
Make sure this target computer is registered in fog
Then setup an image capture as a scheduled task, but before you approve this select the radio button that says to make this a debug capture.

What this will do is boot the FOS client (the fog operating system on the target computer) but it will drop you at a command shell instead of doing the capture right away. Once you are in the FOS shell (on the target computer) then run the showmount -e <ipaddress_of_fog_server> to see from the client perspective what shares are available.

[Miguel Palacios]

@george1421 ok, tell you:

1. I have created a new image (NEW-5420)
2. I launched a capture task with debug capture
3. The laptop has started the FOS shell
4. I have tried to run the showmount -e 10.6.82.2 but the shell response this:

bash: showmount: command not found

?? is this possible?

Also I had tried to run the mount 10.6.82.2:/images /prueba (the path /prueba exist, I had created it this previously) and I get this error:

mount: mounting 10.6.82.2:/images on /prueba failed: Connection time out

[george1421]

@Miguel-Palacios OK, bad advice from me to try to use the showmount command. I thought it was part of the FOS system.

But what is now interesting (I think we are getting close to the problem) while you are in that command shell (by the way, you did great getting here) is to run the following commands.

ip addr show This will tell us if the FOS client is picking up an IP address. I will suspect yes since you are getting this far.

ping 10.6.82.2 to prove you can reach the FOG server from the FOS client.

'ssh root@10.6.82.2` to see if you have full connection to FOG server. (just guessing that ssh is part of the FOS environment).

You already tried the nfs command and received a timeout. If you have success with all of these then we need to focus on nfs on the FOG server. Something is missing then.

[Miguel Palacios]

@george1421 this part seems that is correct… The IP is obtained, the PING is successful and the SSH connection is established.

We guess that the problem is on NFS… right? I saw that rpc.svcgssd, a NFS service, is stopped… is this a problem?

Thank for your help!

[Sebastian Roth]

@Miguel-Palacios There is one thing I noticed:

*/images *(ro,sync,no_wdelay,no_subtree_check,insecure_locks,no_root_squash,insecure,fsid=0)
/images/dev (rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)

Just a typo with the ‘*’?? It should be like this - please check your /etc/exports again:

/images/dev *(rw,async,no_wdelay,no_subtree_check,no_root_squash,insecure,fsid=1)

As well please post the output of sudo iptables -L -n -v on your FOG server here!

[Miguel Palacios]

@Sebastian-Roth Sorry, when I tried write in italics the system “it ate” the asterisk… The /etc/exports seems be correct. Also I show here the output of sudo iptables -L -n -v

[george1421]

@Miguel-Palacios OK, it looks like the FOS client is OK on that target computer.

(step next…)

On the FOG server lets see if we can map itself (also answer Sebastians’s question about iptables).

mount -t nfs <fog_server_ip>:/images /mnt

That should mount the fog servers /images share over the top of the /mnt directory. If you cd /mnt you should see the same content as in the /images folder.

Just so I don’t forget disconnect this mount with umount /mnt.

Have you rebooted this FOG server since you started having a problem? This issue is very strange indeed.

[Sebastian Roth]

I suggest using the nolock option when mounting NFS: mount -t nfs -o nolock <fog_server_ip>:/images /mnt

[Miguel Palacios]

@george1421 and @Sebastian-Roth, the command mount working with itself:

Yes, I have restarted several times, incluse I resintalled the system… U_U

[Sebastian Roth]

@Miguel-Palacios Then try the exact same mount command from the client again! In debug mode that is!

[Miguel Palacios]

@Sebastian-Roth the problem continues…

[Quazz]

@Miguel-Palacios Hi, can you run exportfs -r on your FOG Server, post the output here and try again on the FOGclient? (if it doesn’t work, try exportfs -r a second time)

[Miguel Palacios]

Hi @Quazz , this command executed in the server not show nothing… is it normal?

[Quazz]

@Miguel-Palacios I believe it only gives output if there’s an error, so that’s good news already.

[george1421]

@Miguel-Palacios Please provide the output of sudo iptables -L -n -v

[Miguel Palacios]

@Quazz I have executed other new capture task and the error is the same that my first post
On the other hand, of new, I have executed other new capture task (with debug activate) and I have executed the command mount but this again returned the same error.
In the side server, the command exportfs -r continues without show nothign…

[Quazz]

@Miguel-Palacios What’s the output of rpcinfo -p?