Posts made by mr626

@Sebastian-Roth said in [SVN 7009 FTP errors after image capture: Updating database failed:

@mr626 It’s all over the forums. Try searching…

It means that your FTP credentials are incorrect. This usually happens if you change those but don’t reflect the change in /opt/fog/.fogsettings. When re-running the installer things mess up. Follow this article: https://wiki.fogproject.org/wiki/index.php/Troubleshoot_FTP and don’t forget to get your .fogsettings correct as well!

Hi,

“It’s all over the forums. Try searching…”

I did search before posting, however for some reason the search results were ordered strangely- there is indeed another thread about this exact issue on the same SVN version as me that is a few days old, but it is among other threads that are months or years old (which is I guess why I missed it).

Thanks for your suggestion, even though my FTP credentials haven’t changed I will re-enter them and try again.

Hi guys,

Getting strange FTP errors when trying to capture image. Using Fog SVN 7009. The main part of the image capture goes fine (partclone etc) but when it gets to ‘updating database’ it says ‘failed’.

Then you get the following message repeated a few times:

Error returned: Type: 2, File: /var/www/html/fog/lib/fog/fogftp.class.php, Line 105, Message" ftp_login(): Login incorrect. * Reattempting to update database....failed

The line referenced is:

if (@ftp_login(self::$link,$username,$password) === false) self::ftperror();

This was working fine prior to me updating to SVN 7009. I can’t see any errors in the fog logs or apache logs related to this.

See attached for screenshot

@d4rk3

Ah, I did wonder about that! As mentioned, I’ve not ever played with 8.1 (from a SOE point of view anyway) so I haven’t created 8.1 answer files etc yet.

Edit: going through the thread you mentioned and the links within, I found a Windows 10 specific guide on the same site:

http://theitbros.com/sysprep-windows-10-machine-guide/

I have searched both these forums and the wider internet, but haven’t really come up with much on this topic, other than maybe this:

http://www.tenforums.com/tutorials/3020-windows-10-image-customize-audit-mode-sysprep.html

(I suspect my google search terms are not that good- there has got to be plenty of people out there that have already dealt with this?)

Basically, I’d like to start work on a new standard operating environment image for our organisation. Would be based around Windows 10 Enterprise x64, and activate against our KMS server. Obviously, would be using FOG to capture / deploy the image

Last time I created an image from scratch was for Windows 7 (our currently in-production SOE). So I have a little experience with sysprep, answer files etc.

I guess I am being lazy and hoping there is a good step-by-step guide to preparing a Windows 10 SOE image already!

Suggestions / advise much appreciated. Thanks

@Wayne-Workman

Thanks for all your help. Thanks for the link too, I was hoping there was some more verbose debugging options for the client. I will try this out first thing next week and post back.

@Wayne-Workman

All good points, but sadly I don;t think any of those are the issue either:

-No anti-virus on these hosts yet (other than whatever Windows 10 ships with).
-No IP address conflicts- Fog server has static IP outside of DHCP pool, and hosts get addresses via DHCP.
-Unique hostnames all round (but as mentioned, purely working with IP addresses for now anyway).
-In terms of different ports, I have sort of tried this already. The machines are Intel NUCs (so have both wired and wireless networking options). Have tried both wired and wireless options but there’s no difference.

The machines themselves are identical Intel NUCs- ordered at the same time, same hardware config, even same bios version.

I’ll keep investigating for now.

@Wayne-Workman

Thanks for the reply. I certainly considered the ‘fat-fingered’ possibility, but this is not the case:

1. I am very careful when I enter the IP address, and I double check it
2. As mentioned, I can literally copy the URL from fog.log on the client (that contains the IP address I entered during client setup), paste it into a browser and it works perfectly.

I’m really at a loss, but will keep investigating. Will pull latest trunk and try again.

Edit: problem remains with latest Fog 6333

Hi guys,

Strange issue here. Wasn’t sure to post this here or in developers/bug reports, so apologies if this is the wrong section.

I have a few computers that I am trying to manually install the fog client on (by browsing to the URL http://<my fog server ip>/fog/client and downloading the MSI, and installing it). On most computers it works perfectly, but on a couple I cannot install the client.

*I get a ‘failed to download CA certificate’ pop up message from the installer.
*In fog.log on the client, it has these lines:

PM RegistryHandler 64 bit registry detected
PM Middleware::Communication URL: http://<my fog server ip>/fog/management/other/ca.cert.der
PM Middleware::Communication ERROR: Could not download file
PM Middleware::Communication ERROR: The underlying connection was closed: An unexpected error occurred on a receive

On the fog server, there is nothing in the apache logs (access.log, error.log) to indicate there is a problem.

The clients are on the same subnet as the fog server, and there are no firewalls turned on. From the client I can access / communicate with the fog server perfectly- I can ping it, I can browse to the web interface, and if I past the URL from the fog.log file into a browser I can successfully download the cert.

I can even manually install the cert without a problem.
I am only working with IP addresses at the moment, so there is no DNS involved that could be causing the issue.

Client machines are running Windows 10 Pro x64. Fog server is running on Ubuntu LTS 14.04

Basically, it is like everything works perfectly, but not when the Fog Client installer is doing it (if that makes sense)

Any thoughts?

Ok, I’ve got a workaround that seems to do the trick (for me)

1. Create an image of type ‘single disk, resizable’
2. Capture image
3. In options for the image, select ‘partition 1 only’
4. Deploy

This works for me (tested on one machine so far- Dell optiplex 790, will test on more this afternoon). I should note though- when I created this Windows 7 image originally I made sure that everything was on one partition (ie there isn’t a separate 100mb system partition, just a single partition with everything on it). So the above may not work for everyone.

Hope that helps someone

@Tom-Elliott

Hi, Fog is 6303 as per my first post

@Tom-Elliott

Sorry, was posting from mobile before.

The exact error is as per the OP:

ERROR TRYING TO RESTORE GPT PARTITION TABLES (restorePartitionTablesAndBootLoaders)

And just prior to that, it is attempting to restore a GPT partition table but can’t find one (and has the message du: cannot access ‘/images/<name of image>/d1.grub.mbr’ : no such file or directory)

Happy to run any tests / debug etc if that is helpful. I am re-reading this thread as when I was viewing on mobile I missed a whole bunch of posts somehow, so there’s a few suggestions on here I haven’t tried yet. Will update as I do

Just want to confirm I’m having the same issue as described in the OP.

Have created a new image in Fog web interface of type “single disk, resizable” and re-captured my base image. However on an attempt to image a machine I get the ‘error trying to restore GTP’ error.

Running Fog truck build 6303

Anything I can try as a workaround?

Hi Wayne,

Many thanks for the reply.

To clarify, as part of my troubleshooting process I thought I would try globally disabling and re-enabling the snapin service on the server, via the web interface.

The difference as reported by the client (in fog.log) when I did this was that rather than saying “service disabled for this host” the client correctly reported that the service was globally disabled. And when I then globablly re-enabled snapins on the server, the client went back to reporting that the snapins service was disabled on the host. Hope that makes sense?

Regarding your second suggestion…THANKS! That was exactly the problem! For the problem host, no services had been ticked under service configuration. Almost as soon as I ticked the snapins service, the snapins I had queued started installing.

Follow up question- most of my hosts are still running the legacy client. Does the legacy client ‘know about’ these service configuration settings that are available per host in the trunk version’s web interface? Or do I need to update all of my hosts to the new client (which I will do eventually anyway of course)

Environment Details:

Ubuntu Linux 14.04.2
FOG Trunk version 6207 (just upgraded from 1.2.0, upgrade went perfectly and everything works)
Client: 0.9.11 running on Windows 7 Pro SP1

Problem:

I can’t seem to install snapins. I was getting the exact same problem on 1.2.0 and the legacy client, and I was hoping that upgrading to the trunk version would help (and I needed to upgrade anyway for better Win10 support, so no harm done anyway!).

In fog.log on the client machine, I get the message for the snapin module: “Module is disabled on the host”. I was getting the same message on the host when I was running 1.2.0 and the legacy Fog client (although the message was worded slightly differently- “snapinclient module is disable on this host”

What I tried (on Fog 1.2.0 and legacy client)

Before upgrading to the trunk version, I tried uninstalling and reinstalling the legacy Fog client. I double checked that I was enabling the Snapin module on the host. No change. I also tried turning off the Snapin module globally on the server, and then back on. The client detects this change (and reports it in fog.log) but when it is re-enabled globally on the server I go back to getting the “snapinclient module is disable on this host” message.

What I tried (on Fog trunk 6207 and client 0.9.11)

As mentioned, I did the upgrade to the latest trunk (via SVN method) as I needed to do this anyway for Win 10 support. Upgrade went perfectly, and after upgrade I uninstalled the legacy client on the host, and installed the 0.9.11 one. Unlike the legacy client I didn’t seem to get any options during install to enable / disable certain services though?

I should mention- I previously had no issues with snapins on this exact setup with 1.2.0 and the legacy client. Obviously something changed to cause it to stop working but I’ve got no idea what. On this server I only ever do security updates.

Any help / suggestions are appreciated, thanks. Please advise if there is any extra info you need.

Hi all,

Environment: Fog 1.2.0 running on Ubuntu 14.04.2 server x64. Using latest standard Fog kernel, although I’ve tried the latest unofficial ones too (same behaviour).

Problem 1: PXE bootloop on specific machine.

We have some Lenovo ThinkCentre M72e desktops that I’m trying to register with Fog and image. The PXE booting process seems to go ok. The ‘iPXE initializing’ text appears, and then the text ‘configuring net0’ appears with the mak address of the interface. However, after this the machine reboots (rather than displaying the menu).

I have noticed that the above takes a bit longer than I would have expected: my virtual machines on the same network go through the iPXE initializing and configuring net0 steps in a matter of seconds, but these desktops take a lot longer (and then don’t work anyway).

Is there some timeout value that can be adjusted for this?
Is it worth trying to use .kkpxe rather than the standard (and how do I make this change? In dhcp options?)

Problem 2: occasional freeze during iPXE initializing

We also have some Lenovo X1 Carbon ultrabooks, which use a USB ethernet dongle. These work with Fog fine most of the time, but occasionally when I do a PXE boot with them, the boot process seems to get stuck on the ‘iPXE initalizing’ stage. This seems to be fairly random, and in most cases I can simply reboot the machine and it will then work.

Any thoughts on this one? I’m less concerned about Problem 2 as these machines are only plugged in to ethernet when I’m reimaging them, so staff aren’t impacted.

Thanks in advance for your help

[B]EDIT: Looks like using .kkpxe is the solution to my problem 1. Will this likely have any negative effects for other makes / models that were already working though?[/B]

Thanks for that, I did search the wiki prior to posting but I didn’t see that page!

Ok, noted about 14.04. What about 12.04? I would rather use an LTS release if possible.

EDIT: The title should read ‘fresh’ install, not refresh. Sorry about the brain fart there!

Hello all,

Great work with the new release of Fog guys

I would like to do a completely fresh install of Fog 1.0.1 (will take the opportunity to upgrade the server and install Ubuntu server 14.04 LTS).

However, I would like to migrate some of my data from my exisitng 0.32 install. The main thing would be the hosts, although if I can do the snapins and images as well that would be great.

Is there an automated way to do this? Or is it a case of:

1. transfer the relevant files from old server to new (in the case of snapins and images)
2. export relevant data from current database, and import into new?

Thanks

Sadly, I haven’t had the time to spend on this that I had hoped, sorry

One of the many things I like about FOG is that it is a fairly standard LAMP config- it should work fairly painlessly on just about any distro.

At the moment, our FOG runs on 10.04.3 LTS. Once 0.33 is out, I’ll probably rebuild the whole thing on 12.04 LTS

Fantastic! Many thanks for the reply. I’ll try and get started on this soon