Unable to Mount NFS
-
I don’t think you need to have anything extra installed at OS install time, but it wouldn’t hurt anything either.
I do think a reinstall would help. You don’t have to start from scratch either. Just rerun the installer and it will find your settings file under: /opt/fog/.fogsettings
-
Reinstall cured the dhcp issue and once again I can PXE boot. However the NFS server did not install properly. The package was installed (and came out ok during the check), but was not configured and it returned an error at the end when the services were set. There was also an error about nfs-kernel-server, nfs-common, and rpcbind not being configured correctly due to dependencies or something like that. It flashed by during the install. Unfortunately I did not think to take a picture before restarting (which did not fix the problem). I did attempt to configure NFS using[URL=‘http://forums.debian.net/viewtopic.php?f=16&t=42078’] these directions[/URL], to no avail.
-
Mike,
If you try to install NFS manually, what kinds of errors do you run into:
[code]sudo apt-get install nfs-kernel-server nfs-common rpcbind[/code] -
also, what’s the output of:
[code]cat /etc/exports[/code] -
Here is the output, Tom. The errors encountered during the apt-get mirror those I saw when installing FOG. As for the /etc/exports file, the hostnames are two of the computers on the network (not the server). Before I edited the file there was only one set of arguments for each folder, neither of which had subtree_check, and they were preceed by an * .
[url=“/_imported_xf_attachments/0/491_IMG_20140117_110948.jpg?:”]IMG_20140117_110948.jpg[/url][url=“/_imported_xf_attachments/0/492_IMG_20140117_111033.jpg?:”]IMG_20140117_111033.jpg[/url][url=“/_imported_xf_attachments/0/493_IMG_20140117_111055.jpg?:”]IMG_20140117_111055.jpg[/url][url=“/_imported_xf_attachments/0/494_IMG_20140117_111118.jpg?:”]IMG_20140117_111118.jpg[/url]
-
Not sure if this makes a difference, but the dhcp server was off during the above install so as not to conflict with the campus DHCP server.
-
Try putting this in your exports file:
[code]/images *(ro,sync,no_wdelay,insecure_locks,no_root_squash,insecure,fsid=1)
/images/dev *(rw,sync,no_wdelay,no_root_squash,insecure,fsid=2)[/code]
Then restart rpcbind and nfs-kernel-server:
[code]sudo service rpcbind restart
sudo service nfs-kernel-server restart[/code]Basically the fsid just tells the system in which order the nfs share needs to be mounted. It creates less confusion in the long run.
-
Also, you shouldn’t need net specify the systems. At the point that the system’s loaded, it can’t interpret hostnames anyway (NO DNS SERVER) so it would seem like me it should fail at that point.
-
Sadly that has not helped, although it may make my life easier down the road. nfs-kernel-server and nfs-common are still unrecognized as services when I try and start them (and a reboot hasn’t helped) and the upload still fails when I try and start it. Have we ruled out the configuration files? I’ve attached four which seem relevant.
[url=“/_imported_xf_attachments/0/495_IMG_20140117_115100.jpg?:”]IMG_20140117_115100.jpg[/url][url=“/_imported_xf_attachments/0/496_IMG_20140117_115205.jpg?:”]IMG_20140117_115205.jpg[/url][url=“/_imported_xf_attachments/0/497_IMG_20140117_115252.jpg?:”]IMG_20140117_115252.jpg[/url][url=“/_imported_xf_attachments/0/498_IMG_20140117_115401.jpg?:”]IMG_20140117_115401.jpg[/url]
-
The last one looks a little funny to me. Is that your /etc/exports? Why is it all commented? Maybe add the info I provided above into that one and restart. I might have the service scripts wrong in restart, take a look in /etc/init.d and find the one that deals with nfs. Maybe debian doesn’t use nfs-kernel-server but rather uses just nfs?
[code]sudo service nfs-kernel-server restart
sudo service nfs restart[/code]Maybe do a search for nfs blankly:
[url]http://www.debianhelp.co.uk/nfs.htm[/url]
Has info on nfs and might help.
Of course replace portmap with rpcbind.
-
It turns out you have to add .dpkg-new after both nfs-kernel-server and nfs-common. I was able to restart both although it did not help (also the web interface said the task could not be started). As for the last file it is /usr/share/nfs-kernel.server/conffiles. I have no idea what it does in relation to /etc/exports. Curiously FOG now seems to be able to resolve the host names and it is possible to tell whether they are on or off via the web interface, so some good has come of this.
-
Well I did a fresh install on my home server and encountered no errors whatsoever. Sadly my network would not play ball and I don’t have the energy to deal with it so I can’t say if FOG would totally work. But since none of my research has turned up anything that makes sense it seems the best course of action is to delete the whole VM and start from scratch. Something somewhere may have been corrupted and I’d rather burn the house down then try to root out all of the termites. Fortunately I have a three day weekend to get started.
-
Awesome then. I’m sorry about all the hassle.
-
Not feeling the awesomeness myself, but thank you for all of your help, Tom. Hopefully this will be of use to future users. I suspect something became corrupted during one of the reruns of the installer. It’s too bad there isn’t a script to uninstall the way one installs. If only I had the programming skills I’d write one. Till next time … :D.
-
I can probably make that all happen for you.
The only issue I forsee with it is the deleting of the fog database (if you’re doing a full install of course)
-
Well even Windows programs leave pieces of themselves behind. FOG runs in a virtual machine so I can delete and recreate without much difficulty. I made that decision for this very reason. Part of the issue here is also that I didn’t make a copy of the VM like I should have.
An uninstall script would make a lot of sense for a future release. The effect may be mostly psychological, but it helps to know I can get rid of a program by clicking a file and running a command or two. Also to be perfectly honest after weeks of work I want something guaranteed to be effective. No offense, but I feel a clean slate is badly needed right now and the delete key has yet to fail me :p. Also there is a soft deadline to all of this. The library must go to Windows 7 by April which really means Spring Break (Mid-March) since there will be no other time before then when the library will be nearly empty. So I have to figure all of this out and by then. Except the Dean wants to meet in a week or two and I need something to show him.
-
A thought. Could the names of the NFS services be the root cause of all of this nonsense? The proper terms for my fresh install were nfs-common and nfs-kernel-server as opposed to nfs-common.dpkg-new and nfs-kernel-server.dpkg-new. Portmapper is running, mountd is running, there is no firewall, the config files seem fine. There is no other differences I can spot between my home server install and the one at work.
-
It’s possible, I’m assuming your home server you’re using Debian 7 as well?
-
Yes, that’s why I chose to put it there. It’s not being used for anything else, but is already configured so it was just a matter of downloading and installing FOG. Of course there is a firewall and one of these new Motorola Surfboard modem/router combo deals on the network which will make any PXE booting attempt a pain. So I’ll still create the second VM just in case and rename those files come Tuesday.
-
I have a motorola Surfboard Modem on my network and have yet to have any issues pxe booting. It listens for tftp yes, but only from the isp, not my network as the specifics it’s looking for is unavailable on my network.
