Odd NFS issue
-
What command does fog use to move it? Also, Ctrl + C to drop out of the script, correct?
-
@pugnacious said:
What command does fog use to move it?
I’ll find out… stand by…
Also, Ctrl + C to drop out of the script, correct?
Yes.
-
standing by. I’m taking notes as I run into issues, so perhaps in the future, some other person might benefit from my roadblocks.
-
I mount my /images from an attached array straight into the /images folder and it works without issue. But I’m using ubunutu 14.04 and the drive is formatted as ext4. But regardless I imagine it would still work.
What does your /etc/exports look like?
cat /etc/exports
what does your /etc/fstab say? That’s where the options for mounting it in the os would be
cat /etc/fstab
what are the permissions currently on /images?
ls -l /images
Are you on the latest trunk of FOG 1.2? I think FOG 1.2 had a fogstorage user that needed permissions to the storage nodes and its password needs to be set correctly in the gui storage node settings.
Does the graph of the space usage display on the fog gui homepage?
Are you able to mount the nfs drive elsewhere?
for example if you happen to have an enterprise version of windows install the
“client for NFS” in the “turn windows features on and off” found in the “add or remove programs” control panel.If that client is installed you can mount an NFS share in windows by either
running this command in a run prompt…\fog-server\images
or
In command linemount \\fog-server\images Z: cd /d Z: Z:\>dir images directory listing pops up
Granted @Wayne-Workman is probably right saying you should add it as a separate storage node since its hosted on a different machine and a different OS and all. But some of this might help troubleshoot this if getting it to work is possible through your current configuration
-
@Arrowhead-IT He’s working on making a storage node work on Windows.
-
@pugnacious said:
What command does fog use to move it?
The file that does it is
/var/www/html/fog/service/Post_Stage2.php
-
@pugnacious Looks like it’s two commands, back to back. a rename, and then a put.
if (!$FOGFTP->rename($dest,$src) && !$FOGFTP->put($dest,$src)) throw new Exception(_('Move/rename failed'));
We’re going to do this the easy way.
Edit this file:
/var/www/html/fog/service/Post_Stage2.php
and around line 23 where you see this:
if (!$FOGFTP->rename($dest,$src) && !$FOGFTP->put($dest,$src)) throw new Exception(_('Move/rename failed'));
Change that to:
if (!$FOGFTP->rename($dest,$src) && !$FOGFTP->put($dest,$src)) throw new Exception(_('FAILURE-- Destination: ' . $dest . 'Source: ' . $src));
That will output the destination and source on the screen. This will help us figure out what’s wrong.
-
@pugnacious This is probably a good time to ask… what error are you getting at the end of the upload?
FTP Connection has failed!
orMove/rename failed
-
I just got back to the office. I’m running it again now and i’ll know in a few minutes. Thanks for checking up.
-
It’s moving too fast for me to see the error message. Would a copy of the FTP logs help (edited to remove security details of course)?
From this, it looks like it’s deleting the files, and directory and then quitting instead of moving the files.
Anyway, I won’t be back until monday. I’ll check any reply then.
-
@pugnacious Take a video, upload it to youtube for us to see. From start to finish.
-
Alright, after a full debug capture, stepping through the steps, it shows no message related to FTP at all. I was thinking it was just too fast, but, it’s simply not there. It says Image Uploaded, restoring MBR, Resizing NTFS Volume, clearing ntfs flag, stopping fog status reporter.
Any suggestions on how I can make it report the error?
-
@pugnacious I wonder if Linux FTP commands and Windows FTP commands are the same?
That’d cause an issue.
@Developers Can you help out and give us the FTP commands that are used?
The source directory is
/images/dev/<mac>
and destination is/images/<imageName>
-
I’ve got three machines that I’ve been waiting to image. We need to put those in production and the work is already done creating the images for them. We just need to do a quick clone. I’m thinking it’s either the ftp commands or possibly the NTFS file system on the NFS share. To bypass the problem, i threw together a system and stashed it under my desk running CentOS 7 and with a decently sized hard drive. Now i’m running into another issue installing fog to the machine. It’s running trunk (svn) and it’s freezing at “Downloading inits, kernels, and the fog client”.
I did notice that the svn build was behind what i’m using on the other server. I did use the command to update the svn, and it said it was the same build. I’m going to try and download a fresh copy through git instead as that seems to be the most updated.
-
@pugnacious git and svn have identical files, both are valid. hanging at the downloading inits and such, that could be a DNS problem, or it could just be busy right now, or maybe a slow internet connection. Give it time, be patient. At my work, that normally takes about 30 seconds.
Just to get yourself going, you can manually move the image on the other server. In the web interface, if you look at the image definition, it’ll have a “Image Path” field. Rename your image to that, and just move it from /images/dev/ to /images and then try to deploy the image to those computers you need to image.
-
@pugnacious Did you disable the firewall and SELinux on the CentOS one? From what I’ve read, they can cause a host of issues on that OS in relation to FOG.
-
@Quazz I’d actually recommend @jbob 's firewalld configuration. I’ve had zero issues with it.
https://forums.fogproject.org/topic/6162/firewall-configuration
for service in http https tftp ftp mysql nfs mountd rpc-bind proxy-dhcp samba; do firewall-cmd --permanent --zone=public --add-service=$service; done systemctl restart firewalld.service
but yes, SELinux needs set to permissive, instead of being enforcing.
vi /etc/selinux/config
Instructions on using Vi:
https://wiki.fogproject.org/wiki/index.php?title=ViChange
enforcing
topermissive
in that file.
Then to change the live setting, issue:setenforce 0
-
Alright, I’ve swapped out the VM to a full machine. Got it working on the latest GIT version, and found a rogue dhcp server causing issues. Now though, after taking an image of a reference machine, it gives me the error while trying to redeploy the same image to the same machine, Image Store Corrupt Unable to locate MBR.
The machine is a single disk resizable running windows 7. Any thoughts?
-
@pugnacious what version are you running?
-
Version:
61226128
bzImage Version: 4.4.0
bzImage32 Version: 4.4.0Edit: Just updated to check. Still no go. Rebuilding that machine from scratch and i’ll take a new reference image i guess.