Posts made by george1421

@Kiweegie I will confirm on Centos 6.5 to save you time. I have a template already setup in ESXi to spin up a new instance. I just haven’t do this test yet. I’ll have a go-no answer for you in about 30 minutes.

@ITSolutions I created a new tutorial that requires the reader to create a new ipxe boot kernel to do as you want. It did work in my environment here is a link to the tutorial if you want to give it a try. https://forums.fogproject.org/topic/6400/usb-boot-uefi-client-into-fog-menu-harder-way

WARNING: this tutorial is no longer relevant since the rom-o-matic site has been shutdown since Dec 2019. These instructions are no longer valid and can be followed. I'm leaving these instructions here for historical reasons. No additional support on this topic can be provided since they depend on an external web site working

You may has why do we need this?

UEFI PXE booting is a bit different than BIOS based PXE booting. Some of the early UEFI systems like the Dell Latitude e6420 and the OptiPlex 790 do not support PXE booting in UEFI mode. But through testing they do support USB booting in PXE mode. So knowing this I’m going to create the harder solution (see easier solution if this method is too complicated for your needs) to PXE boot these devices in UEFI mode. This method we will create a new efi boot kernel to install on our usb flash drive to boot into the FOG environment. Building linux kernels are not extremely difficult using the rom-o-matic, but because of the number of options you can choose, you can run into problems if not created correctly. I can say for sure this method works with the previous mentioned systems. YMMV with other hardware platforms.

The process steps are not that hard, its a bit like giving your self a root canal without medicine. It will hurt like crazy while you are doing it, but once its over you will have the satisfaction of not wanting to ever do it again <jk>. Its not really that hard, the rom-o-matic does all of the work.

You will need to acquire these things.

1. A 2GB (min) flash drive
2. A UEFI pxe boot image we will create from the rom-o-matic web site

iPXE UEFI boot kernel creation process

1. From a browser access the rom-o-matic web site at https://rom-o-matic.eu/
2. Select Advanced, for experienced users radio button
3. In the drop down list for the output format select EFI PXE bootstrap 64-bit (.efi) selection.
4. Ensure that NIC type is set to all-drivers (default)
   Since there are a ton of settings that could be changed here, I'm only going to post the changes from the default settings for the PXE boot. HINT: You may want to search in your browser for the variable names I used below to locate the exact setting in question.
5. Section: Download protocols.
   DOWNLOAD_PROTO_HTTPS = checked
   DOWNLOAD_PROTO_FTP = checked
   DOWNLOAD_PROTO_NFS = checked
6. Section: SAN boot protocols
   All checked in this section except HTTP_ENC_PEERDIST = unchecked
7. Section: Image types
   All unchecked except IMAGE_PNG = checked
8. Section: Command-line commands to include
   All checked except IWMGMT_CMD = unchecked, PXE_CMD = unchecked, PROFSTAT_CMD, = unchecked
9. Section: Console options
   All unchecked except CONSOLE_FRAMEBUFFER = checked
10. Section: Embedded script add in the following script

#!ipxe

set fogserver 192.168.1.88

isset ${net0/mac} && dhcp net0 || goto dhcpnet1
echo Received DHCP answer on interface net0 && goto netboot

:dhcpnet1
isset ${net1/mac} && dhcp net1 || goto dhcperror
echo Received DHCP answer on interface net1 && goto netboot

:dhcperror
prompt --key s --timeout 10000 DHCP Failed, hit 's' for the iPXE shell; reboot in 10 seconds && shell || reboot

:netboot
cpuid --ext 29 && set arch x86_64 || set arch i386
params
param mac0 ${net0/mac}
param arch ${arch}
param platform ${platform}
param product ${product}
param ipxever ${version}
isset ${net1/mac} && param mac1 ${net1/mac} || goto bootme

:bootme
chain http://${fogserver}/fog/service/ipxe/boot.php##params

Be sure to change the ip address above for next-server to the IP address of your FOG server. You must use an IP address and not the conical name of your fog server.

11. Section: Which revision should be set to Master (default)
12. Press the Proceed >> button. After a bit you will be prompted to download a file called **ipxe.efi ** be sure to save it. We will need this file in the section below.

Boot drive creation process

1. Insert your flash drive into a Windows based computer and format it with FAT32 disk format
2. On that flash drive create a folder called EFI
3. On that same flash drive create a folder called BOOT in the EFI folder creating this path “x:\EFI\BOOT”. Note: I have not tested if case is important or not, I used upper case for everything and it worked. That is as far as I tested.
4. Copy the ipxe.efi kernel created in the previous section to the flash drive in the EFI\BOOT folder. That file MUST BE RENAMED to bootx64.efi (note the case difference. I did not test to see if case is important)
5. At this point remove the usb thumb drive from the build up computer and insert the drive into a target computer
6. Power on the target computer and press F10 or F12 (depending on the mfg) to call up the EFI boot menu.
7. Select the USB boot device under the EFI section of the EFI menu
8. You should see the iPXE boot banner and then after about 30 seconds it should be prompted for the IP address of your FOG server. Key in the IP ADDRESS of your FOG server and press Enter.
9. At this point you should boot into the FOG iPXE menu.

Using this (easy way) method, no.

What we would need to do is compile a new efi kernel as I outlined in the tutorial for the BIOS usb boot [ https://forums.fogproject.org/topic/6347/usb-boot-bios-client-into-fog-menu ]. I attempted to do the same workflow for the EFI boot as BIOS but the kernel didn’t work. I decided to take the “easy way” and rely on the work done by the FOG developers.

Now doing what you want is totally possible, if we can get the kernel configurations from the FOG developers, because the rom-o-matic can be used to compile custom kernels. We just need to know the options to select. This also assumes that the FOG developers haven’t patch the ipxe kernels in their environment.

To add a little background information to this. The error noted in the picture is related to configuring the nic. From the ipxe error page this is stated.

This error indicates that iPXE timed out while attempting to autoconfigure a network adapter.

In almost all cases the method used for autoconfiguration will be DHCP, and you should read the information relating to DHCP timeout. 

ref: http://ipxe.org/err/040ee1
While working on a project over the holiday. With a system booting and using dnsmasq I saw the device disconnect and reconnect to the switch 3 times during the boot into the ipxe menu. The last link drop was just after printing Configuring (ne0 …). So that makes me think that this is a spanning tree issue (which can be resolved with portfast or RSTP).

@Wayne-Workman said:

@george1421 a full install is fine. It’s the storage node install that fails. I had the exact same issue with a new CentOS 7 install

Let me make sure i understand this. The issue is with the storage node install, right? If so I can spin up a new centos 6.5 instance in about 10 minutes. I can confirm this with a clean install if needed. But I’m not seeing this as an issue unless the FOG installer is bringing in other repositories. That also doesn’t explain how/who php5.6 is install on this target computer. Either way I can confirm this on a clean install.

@Wayne-Workman Very nice job. I’ll be sure to test this out on both Centos 7 and 6.5 after the holiday. But overall this looks great. It checks all of the boxes on the wish list.

@Kiweegie The error message says that you have php 5.6 installed on this system, additionally you have other repositories enabled for this client. I see epel and remi, rpmforge all enabled. This is the cause of the issue with this box you have conflicting packages.

This part is a guess (although probably spot on) you are trying to install this RPM php-fpm which is for php 5.36 which is conflicting with 5.6. I would assume the right package for 5.6 is php56-fpm. This is not a FOG problem but a system owner issue (Merry Christmas). You probably have other stuff loaded on this FOG server that needed php5.6, which would explain the other repositories being loaded.

I have FOG trunk running on Centos 6.5 without issue, including the trunk upgrades. My dev FOG environment is running Centos 7 without issue too. But in both cases I only have FOG running on these servers.

@Sebastian-Roth I would tend to agree here with the network being (partially) at fault. I have seen spanning tree cause an issue because it takes too long to move the port into a forwarding state once the link up signal has been sent. By then OS has already given up on the link thinking that it is down. This will generally cause dhcp issues FOG or not, for the first minute or two the port link is up. They created RSTP and portfast to address the slow 30-60 seconds wait for STP topology check.

One quick way to check this is to turn off spanning tree all together on that port where the device is. (ONLY TURN IT OFF to test if spanning tree is your problem or not. Do not leave the port in this state if it is in the user domain)

@Wayne-Workman

Some other things I just thought of while walking about.

How does dhcp lease time come into play in regards to script execution. Especially if this will be a cron job. Consider what is a shortest duration lease and a typical duration lease time.

should the physical location of this script be location in a fog related directory like /opt/fog/utils?

Is there any way to trap an interface going up/down and then to call your script? hint: http://www.cyberciti.biz/tips/how-do-i-run-firewall-script-as-soon-as-eth0-interface-brings-up.html

Does the interface go down and then up on a dhcp renewal.

We would assume that the path statement points correctly for bash to run without full paths on the called applications especially if they are system type applications.

@Wayne-Workman This is going to be a critical review of the code. Don’t take it personally (my guys hate me when I do this type of review, but the finished product is generally better).

Some lines use the full path to the application and some only use the application name. (i.e. /usr/share/awk)

Since you are using the full path to the application you assume that is the proper location. I just checked on my ubuntu based system awk is in /usr/bin/awk under Centos 7 its again /usr/bin/awk but in your script it references /usr/share/awk which I assume would cause your script to fail on these systems.

What would happen if interface or fogsettingsIP returned nothing instead of a value what would happen to your script?

The same question if the mysql credentials were not returned.

In sudburr’s post also references a wol host field to be updated, is this important?

What happens if config.class.php.old already exists when the script is run a second or third time. Will anything important be lost?

Do any running fog services maintain a cached copy of the server’s IP address or do the fog services need to be stopped when you update these settings to make them stick?

I agree this is great! I’m glad we have the community involvement in the FOG community to help everyone solve these basic problems. Well done @sudburr !!

The only thing that jumps out that may be an issue in the script is that for the Centos 7 OS the ethernet adapters are no longer just eth0, eth1 and so on. The Intel NUC I’m working on the default ethernet adapter is enp3s0.

I just checked and this value interface="enp3s0" can be extracted from the /opt/fog/.fogsettings file. This would then tie into what interface FOG is using too. There may be a case where there is a system with one or more network interfaces that the script would have to take into account.

@Wayne-Workman dhcp will work, the issue is that some dhcp server’s don’t have reservations. (yes this could be for a mobile deployment server, so we can’t garanty we will have access to any settings on the dhcp server). Also it is generally good practice to set any device that offers a service to the network a static address. Will FOG work with dhcp sure in a stable environment. But it is a bit of a PITA (as I found out) if you need to change the IP address later. Fix it before fog is loaded is the best recommendation here.

As far for dnsmasq, yes that is needed. I have another Tutorial on setting that up under Centos 7

Some systems no longer come with CD/DVD roms, to install Centos 7 on these systems we must transfer the downloaded ISO image to a usb flash (thumb) drive.

1. Download Cento 7 from your favorite mirror
2. Create a bootable Centos 7 image on the USB flash drive. If you are windows centric user, follow the reference link below that uses Win32DiskImager utility. Since I’m using a Zorin (Ubuntu variant) based laptop I’ll just use the dd command.
   ref: http://www.sysadminguide.net/how-to-create-bootable-usb-key-for-centos-7-installation/
3. Insert the flash drive into your linux computer.
   To determine where the flash drive is mounted issue the following command: lsblk
   Your output should look similar to this:

NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda      8:0    0 298.1G  0 disk 
├─sda1   8:1    0 294.3G  0 part /
├─sda2   8:2    0     1K  0 part 
└─sda5   8:5    0   3.8G  0 part [SWAP]
sdb      8:16   1    15G  0 disk 
└─sdb1   8:17   1    15G  0 part /media/johndoe/BEA0-A4EA

4. As you can see from the printout my 16GB flash drive is device sdb and its mounted on /media/johndoe/BEA0-A4EA. Armed with this information, we’ll need to first unmount (not remove from the computer) the flash drive before we use the dd command to transfer the image. To unmount the flash drive issue the following
   umount /media/johndoe/BEA0-A4EA
5. Once the drive has been unmounted I use the dd command to write the iso image to the flash drive. The syntax for dd is dd bs=4M if=CentOS-7-x86_64-DVD-1511.iso of=/dev/sdX where X is the device number of the usb flash drive. Get this right or you could overwrite something you might need later. Using the information from the lsblk command I know my flash drive is /dev/sdb so the proper dd command to write the iso image to my flash drive is: sudo dd bs=4M if=CentOS-7-x86_64-DVD-1511.iso of=/dev/sdb Depending on how fast the flash drive is it may take several minutes to write the iso image to the flash drive. On my laptop this was the output from dd.

1032+1 records in
1032+1 records out
4329570304 bytes (4.3 GB) copied, 521.836 s, 8.3 MB/s

6. Remove and reinsert the flash drive into your computer. Your computer should properly mount this flash drive.
7. Viewing the contents of the flash drive should be successful.
8. Done with the copy, now move the Centos 7 bootable flash drive to the target computer.

The installation of Centos 7 and FOG on the Intel NUC is really no different than on any other physical machine. I’m outlining the steps required because Centos 7’s default install behavior is not what is needed for a typical FOG install. This default behavior is to take the physical disk and give 50GiB to the root (logical) partition, 500MiB to the boot partition and the remainder of the physical disk to the /home partition. This is donating a large portion of the disk to the /home partition where currently FOG stores all files in the root partition.

Install Centos 7

1. Insert your boot media into the NUC’s USB port. If you haven’t converted the ISO image to a bootable flash drive check out this link: https://forums.fogproject.org/topic/6386/copy-centos-7-iso-images-to-usb-flash-drive
2. Boot the NUC
3. Press F10 (F12 on other manufacturers hardware) to enter the bios boot menu
4. Select your boot media (USB)
5. When the kernel boots, select Install Centos 7
6. When the gui loads, enter your language choice and press Done
7. On the next page everything should turn ready after a few seconds execpt the install destination.
8. Select the Install Destination
9. Select the disk to install Centos 7 on. When the disk is selected there will be a check mark on the disk icon.
10. In the Additional Options select I will configure disk. (from memory will need to confirm actual text)
11. For this install if there are any existing partitions remove them. We will create a new layout specifically for our fog server.
    Create this hard disk partition format using LVM

/boot 500 MiB
swap 8192 MiB  (actually should be based on the amount of RAM installed. On my system I have 8GB of ram so I'm creating a 8GiB swap)
/ 24 GiB
/opt XXX GiB (should be remainder of disk or array. This is where we will eventually store all of our images and snapins)

12. Press Done
13. Press Begin Installation
14. Enter root’s password
15. reboot when installation is done.

Configure Centos 7 before installing FOG.

1. Log into the console of the NUC using root and root’s password
2. The first order is to setup the network interface.
   cd /etc/sysconfig/network-scripts/
3. Edit the ethernet adapter’s config file. Note that the name of the ethernet adapter on the NUC is enp3s0
   vi ifcfg-enp3s0
4. Set the network adapter to initialize on boot by changing the value to ONBOOT=yes
5. You MUST enter a static IP address for your FOG server or bad things will happen. Changing the IP address after FOG is installed is problematic at best. Insert the following text into the ethernet adapter config file. Make the values appropriate for your network.

IPADDR=192.168.1.15
NETMASK=255.255.255.0
GATEWAY=192.168.1.1
DNS1=8.8.8.8

6. Save and exit from your text editor

7. Restart the network subsystem to take on these new settings.
   systemctl restart network

8. Ensure that Centos 7 is at the latest release with all of the fixes applied before you start.
   yum upgrade -y

9. Install a few needed extra packages
   yum install wget dnsmasq subversion -y

10. Disable the linux firewall
    systemctl disable firewalld.service

11. Disable selinux
    vi /etc/selinux/config
    SELINUX=disabled

12. Reboot the system to take on these new settings
    reboot

13. Log back in as root

14. Install the latest SVN trunk (or the latest stable release here)
    svn co https://svn.code.sf.net/p/freeghost/code/trunk /opt/fog/fog_trunk

15. Change to the FOG installer’s directory
    cd /opt/fog/fog_trunk/bin

16. Run the installer ./installfog.sh

17. Install FOG using the settings most appropriate for your environment.

Some things that haven been addressed yet is the default location that fog uses to store the captured image files is /images. This happens to be on the root partition where we allocated 24GiB of storage. We set aside space on the /opt partition for all of the big files. During the FOG install it will ask you where the image directory should be, I would recommend that you place it in /opt/fog/images and not on the root partition. The reason being is that historically bad things happen if you use up all of the space on root partitions under unix. Usually you will break the OS if you fill up the root partition to the point where you will need to reinstall the OS to fix it. So (IMO) its bad practice to store anything on the root partition that might cause it to fill up, hence the reason for creating the /opt partition during OS setup. We can fill up the /opt partition with captured images without risk of taking down the system.

I have more to add to this. I did a clean install of fog this AM and it again failed at setting up storage. I selected all default settings and no monkey business with the files location and it still failed. After checking to ensure I didn’t do anything I reran the installer again with the same default values and it installed correctly.

Additionally I discovered an issue that the ftp (OS) account was setup with the correct password (as indicated by the install script), the FOG_TFTP_FTP_PASSWORD value was set to the password given to the ftp user on the first install attempt not the second.

@Tom-Elliott The probably the easiest solution is to create a procedure (wiki page) that if you need to update your fog server you need to do:

1. Update the settings in the /opt/fog/.fogsettings
2. Rerun the installer
3. Update the IP address for the storage node on the FOG system where you changed the IP address
4. Update the IP address on a any master storage node that may reference this FOG server
5. Update the FOG_WEB_HOST value
6. update the FOG_TFTP_HOST value

I can I’ll marked this as addressed (Solved), since its not a technical solution but a procedural one that resolves the issue.
[Edit] I would, but it appears I can no longer mark topics as solved[/Edit]

I also understand what you are saying, but what would the impact be if the installer just blindly updated the current values (what ever they are) to the settings found in the .fogsettings file?

However this could be also addressed with a wiki page. Once the settings were found the fog server did work as intended at the new IP address. In the end this was my own fault for being lazy and letting dhcp setup the OS and not remembering to put it at the final IP address before fog was installed.

Once fog is installed it is difficult to change the IP address of the fog server.

I updated the /opt/fog/.fogsettings with the new IP address and then reran the installer script.

The first issue I ran into was the background wouldn’t load for the iPXE menu. This was addressed by updating the fog setting FOG_WEB_HOST to the new IP address.

The second issue I ran into was not being able to connect to the NFS share. This was corrected by updating the default storage node, default member, IP address.

Also I see FOG_TFTP_HOST setting still points to the old IP address

I would be great if the installer script would check these values to see if they were different then update them to reflect the new setting in the .fogsettings file.

On a new install of the fog trunk the FOG_DEFAULT_LOCALE is set to an asian default value (sorry can’t tell if Japanese or Chinese) Since this value is at list index 0, I might think the actual value isn’t initialized in the database until I set it from the web GUI.