Posts made by albion

@george1421 Even though the machine “seems” to be connected to the domain after reimage, I always remove and readd the machine to the domain.

On my current “problem” I had to use the manual complete reset method to get Windows Update working again. And I’m still not sure because it’s “Checking for Updates” on 2 years worth of updates.

@Wayne-Workman: Fog version 1.2.0. Linux CentOS 6.7 server, Windows 7 SP1 Workstation clients.

@MRCUR: I am using WSUS. But, points at george1421. Each Workstation image will only be restored to that Workstation with a new same model hard drive. The impression I always got was that sysprep wasn’t necessary in my case.

I basically use fog as a workstation backup solution. After I install a PC, updates and software for a user I use fog to take an image of that machine. One problem that I see reoccuring after a re-image is a failure to run Windows Updates. Is there something I need to do before I try to run Windows updates on the re-imaged machine?

@Wayne-Workman

Model: Whitebox ASrock MB, Core i5 CPU
Winodws 7 Sp1 x64
MBR
Image: I installed from a Microsoft SA Win7sp1x64 downloaded ISO.
When I hit DEL BIOS starts in UEFI Setup Utility.
The SSD was operating in AHCI mode. The new HDD is the same.

@sudburr

I’m not taking images for distribution. I install a machine, install the OS, and install that particular users software. I then take an image of the machine in case there’s a virus or hard drive crash. I haven’t gotten to distributions yet, although that is my next step.

Recently I had a primary hard drive crash on a users machine. I immediately installed a new HDD from stock. I went to my Fog server and set download on the single drive resizable image I had taken when I installed that machine. The download worked but I was left with the old “Windows failed to start” error screen. I tried a repair from the Windows 7 disk but that didn’t help. Eventually I just reinstalled the machine. After reading some more I came to the conclusion that I did not prepare the workstation properly for image upload.

I looked all over for a howto describing the preparation of a workstation for image upload for Fog version 1.2.0, but all I seem to be able to find are docs mentioning older versions. Some say run fogprep, some sysprep, etc… Can someone post a detailed howto on preparing a workstation for image collection for version 1.2.0?

Thanks.

I know this is a bit old but I just wanted to send off a “Fixed my problem” message. I had moved my /images from a directory on the root partition to a new larger partition mounted to /images. When I made the cp -a it did not, for some odd reason, copy either of the .mntcheck files in /images or /images/dev. After creating and setting permissions on those two files everything started to work again. Thanks much!!

[quote=“Tom Elliott, post: 35955, member: 7271”]I think I know what the problem is with it.

It’s not DNS related, though your DNS may have an issue.

It’s related to the fact that we don’t install ClamAV on Centos/Fedora/Redhat OS’s anymore and I’m thinking the install function had a file that get’s requested. As the clamav isn’t installed, I’m thinking some directories aren’t copied over.[/quote]

Basically you’re saying Ubuntu over CentOS. I suppose, although a neat and possibly useful tool, I can live without it for now. Then again, maybe I’ll try an Ubuntu install.

Thanks for all your help.
-Craig

The main.cvd download got to about 10% then failed on a full disk. “getfile: Can’t write 1440 bytes to /usr/share/clamav/…”

The firewall was disabled during the FOG install as instructed. I can freshclam once I add /tmp/resolv.conf in debug mode, although the download is rather slow. 5 minutes and only at 3% of the main.cvd download. I have 10meg, so I am assuming the download server is not in the US.

[quote=“Tom Elliott, post: 35910, member: 7271”]/etc/resolv.conf should have:
nameserver <IP Of Router Gateway in house, or direct dns entries (208.67.222.222, 8.8.4.4, etc…)

Then check that resolving works by:
ping google.com

if it doesn’t, maybe there’s a proxy in between?[/quote]

This is my procedure for changing that file.

0. cd /var/www/html/fog/services/ipxe
1. xz --decompress init.xz
2. mount init ./tmp -o loop -rw
3. cd ./tmp/tmp
4. touch resolv.conf
5. echo “nameserver 8.8.8.8” > resolv.conf
6. echo “nameserver 8.8.4.4” >> resolv.conf
7. cd …/…
8. umount ./tmp
9. xz -z -9 -C crc32 init

I then told my test bench to reboot in debug mode through my Fog server. The machine shutdown windows and rebooted me to a Linux command prompt. Although the link for resolv.conf was in /etc, the file I created in /tmp didn’t exist. When I tried to ping google.com I got the response “ping: bad address ‘google.com’”. After I ran though stops 4 to 6 from above while in the debug console my “ping google.com” then ran successfully. Next I tried rebooting the debug console to see if the resolv.conf I created in the debug console was still there… no luck.

-Craig

[quote=“G0dzilla, post: 35883, member: 1692”]After the resolv.conf file is added to the init.xz it still needs entries in it for DNS to work.
This can be done manually or with a program like udhcpc.
This can all be done in a debug task - followed by the command “freshclam” to see if clamav can update. This may not work still, but should get the DNS working at least.

I believe the database files are left out to save space in the init.xz .
The init.xz file is currently limited to 57MB I think.
Here is info on increasing it - [url]http://fogproject.org/forum/threads/change-initrd-size.3920/[/url]

I think a more ideal solution for Clam with FOG would be for the clients to download the database files from the local fog server instead of going out on the internet for them. It would involve a bit more manual customizing for each server tho. I think virus scanning with FOG is just a cool feature - but isn’t really important.

You can always try Trinity Rescue Kit for scanning, that has 4-5 scanners built in and can be PXE booted.
There is also a script available to run all the scanners.
[url]http://code.remyservices.net/trk_scripts/wiki/Home[/url]

remy_virusscan - Offers the ability to update all scanners at once and then scan all attached drives with the updated scanners in a silent mode. Scanners include Avast!, BDE, CHKRootkit, ClamAV, F-Prot, RKHunter and Vexira.[/quote]

I know a little bit about servers. I set the nameserver entry to 8.8.8.8 (Google name server). I also tried my two internal DNS servers. Neither Google nor my internal servers worked. I didn’t though add search or domain directives to resolv.conf, are they necessary for this implementation?

It’s possible to PXE boot from two different systems? i.e. Fog and AV Offline Scan.

[quote=“G0dzilla, post: 35840, member: 1692”]Here is some info I have found that I think may be related.

From what I can see, there is no /etc/resolv.conf file in the init.xz, and no default program to add entries to resolv.conf [dhclient, wicd etc, I don’t know what program is currently used to get a DHCP address] .

Inside the init.xz there is a sym link;

/etc/resolv.conf -> /tmp/resolv.conf

/tmp/:
total 1.0K
ldconfig

/tmp/resolv.conf does not exist.

Things that may help testing/fixing;

You can edit the init.xz and add the /etc/resolv.conf [or /tmp/resolv.conf].
You can run udhcpc to automatically add DNS servers to resolv.conf[/quote]

After finally figuring out how to uncompress and mount the init file I was able to add the resolv.conf. That made no difference. But I also noticed that the error message is telling me there are no database files in /usr/share/clamav. I check that out as well and noticed there are no database files there. When I try to add the files from another install I am met with a “No space left on device” error.

[quote=“Tom Elliott, post: 35802, member: 7271”]It’s most likely that the DNS isn’t being sent for the clamav task, I’ll see if I can get around to tracking the bug.[/quote]

I guess I was kind of assuming that I didn’t have something setup correctly. I wasn’t at all thinking it was bug. Thanks.

Yes this is [quote=“G0dzilla, post: 35743, member: 1692”]Is this the error ?

[ATTACH=full]1309[/ATTACH]

I’m also running Centos 6.5 + FOG 1.2.0.

The init.xz [i think its this file] doesn’t have a database in it to use already, and it cant go out and download one hence the error. Need a file for clamav to use in the init.xz, or verify your client has connectivity to the internet, can test this with a debug task.

I’ll look more into it later.[/quote]

That’s the error alright. I’ve tried installing the latest clamAV using yum, but that didn’t do anything. It probably screwed something up. lol

Sorry to bother everyone. I’ve been all over the web looking for help with ClamAV and Fog. Although the fog_1.2.0/installation.txt claims a CentOS ClamAV Wiki page, I can find no such thing. So I am hoping someone here has been successful in getting ClamAV to work with Fog 1.2.0 under CentOS 6.

Whenever I set the ClamAV task in fog manager the machine boots properly into PXE and tries to start the ClamAV program. I get four warnings complaining about the lack of access to DNS and ClamAV update sites. It gives those warnings twice and then an error about an inability to find a file or directory. The messages fly by so fast that I had to restart the task 5 times just to remember what I wrote above.

I would surely appreciate any help with this.

-Craig