Web GUI with problems in FOG 1.0.1

[Baite]

Upgraded FOG to 1.0.1 in CentOS 6.5.
Updated undionly.kpxe to the one named undionly.kpxe.INTEL
I have everything working, except for the WEB GUI.
For example, in Firefox and Chrome, after entering Host Management, I cannot access any of the links in Host Menu (i click on the links, but nothing happens).
In IE 11 it is working.

I need to say that I have 2 NIC cards on my FOG server, each in a differente VLAN. The IP configured in FOG is 10.1.8.1, but I access the Web interface with another IP, in another network (10.1.2.x).

[Tom Elliott]

So which interface is the WEB GUI accessible from? Are the VLAN’s able to communicate across each other? If they’re not able to, this is probably why you’re unable to see the GUI.

Which one is hosting TFTP? What are the error’s you’re seeing with that? You wrote on my wall that things aren’t working, but we need more information. Assuming the problem is linked to the fact that you have two separate VLAN’s that you want communicating with each other but aren’t, I’d say start with your DHCP server. I’m going to guess that you have a Singular DHCP server doling out IP address to each of the respective VLAN’s within your control? If this is the case, then you need to specify the option’s 66/67 on the network. Also, check your iptables settings and make sure firewall is turned off.

[Baite]

[quote=“Tom Elliott, post: 28235, member: 7271”]So which interface is the WEB GUI accessible from? Are the VLAN’s able to communicate across each other? If they’re not able to, this is probably why you’re unable to see the GUI.

Which one is hosting TFTP? What are the error’s you’re seeing with that? You wrote on my wall that things aren’t working, but we need more information. Assuming the problem is linked to the fact that you have two separate VLAN’s that you want communicating with each other but aren’t, I’d say start with your DHCP server. I’m going to guess that you have a Singular DHCP server doling out IP address to each of the respective VLAN’s within your control? If this is the case, then you need to specify the option’s 66/67 on the network. Also, check your iptables settings and make sure firewall is turned off.[/quote]

Tom:
I’m sorry to have posted that information in the wrong place. Regarding that subject, both VLANs don’t see each other, but I managed to have things working again (as it was working prior to version 1.x.x).
I have been using Fog 0.32 in my school for a long time. But now we have 35 new machines with UEFI BIOS, and we weren’t able to deploy them using 0.32. So, we decided to upgrade to version 1.x.x
Right now, I am using latest version (1.1.0 which, by the way, says it’s not the last version in Fog Settings), and having some dificulties.
The first one is that I get PXE-E53: No boot filename received (and option 66 - nextserver 10.1.81 - and 67 - undionly.kpxe - are configured in my CISCO switch (which is the DHCP server for this VLAN).
Last, I read in some forum, about compiling specific versions of undionly.kpxe ([url]http://fogproject.org/forum/threads/realtek-8111-8168-undionly-kpxe-hangs-on-initialising-devices.10453/[/url]), but I have no idea of the process and how to achieve this… for example, what is chainloading and why it’s necessary. I read at “[url]http://etherboot.org/wiki/pxechaining[/url]” something, and also you article at “[url]http://fogproject.org/forum/threads/build-ipxe-from-source.10152/[/url]”.
My older machines all are a HP dc7900p with Intel PXE, but my new machines are ASUS B85M-G onboard NIC from Realtek. And I need desperatly to have things working, for I’m the only one who deploys and handles about 400 computers at my school, although my main job is teaching.
Sorry for taking your time… I want to be able to do it alone withouth bothering you further, but I need some lights first, if you be kind enough. I am willing to publish all the work, so others can repeat it and resolve their problems to.

Thank you, once again, for this incredible work you’ve done. God bless you.

[Junkhacker]

UEFI needs to be disabled. at present, only legacy boot works.

[Baite]

I’ve done that… but the old PCs (non-UEFI) are unable to receive the undionly.kpxe file now.
In older machines, I get:
[INDENT=1]PXE-E53: No boot filename receive[/INDENT]
[INDENT=1]PXE-M0F: exiting Intel Boot Agent[/INDENT]
In the new machines, with Legacy mode enabled, I get (the first time I booted):
[INDENT=1]PXE->EB: !PXE at 99C6:0070, entry point at 99C6:0100[/INDENT]
[INDENT=2]UNDI code segment 99C6:3ABC, data segment 0D7B:C4B0 (565-630kB)[/INDENT]
[INDENT=2]UNDI device is PCI 03:00.0, type DIX+802.3[/INDENT]
[INDENT=2]565kB free base memory after PXE unload[/INDENT]
[INDENT=1]iPXE initializing devices…ok[/INDENT]
[INDENT=1] [/INDENT]
[INDENT=1]iPXE 1.0.0+ (7627f) – Open Source Network Boot Firmware - [url]http://ipxe.org[/url][/INDENT]
[INDENT=1]Features: FTP HTTP HTTPS iSCSI DNS TFTP VLAN AoE bxImage ELF MBOOT PXE PXEXT Menu[/INDENT]
[INDENT=1]Configuring (net0 e0:3f:49:46:8d:e4)…"[/INDENT]
Then, although a bunch of messages with error appear (see attached photo), I could acces Fog Menu and,for example, register host (full).
In GUI I can see the new host, added it to a group, and in basic tasks, ordered image to upload.
Now, it keeps looping in boot, appearing everything until “iPXE initializing devices…”. Deleted task, but problem remains.

With all this in mind, I think it may have smthing to do with some incompatibility of the NIC with undionly.kpxe… what can I test more?

[url=“/_imported_xf_attachments/0/829_2014-05-27 14.55.46.jpg?:”]2014-05-27 14.55.46.jpg[/url]

[Baite]

Just to send other photos of the boot process of the new machines…

[url=“/_imported_xf_attachments/0/830_2014-05-27 14.49.10.jpg?:”]2014-05-27 14.49.10.jpg[/url][url=“/_imported_xf_attachments/0/831_2014-05-27 14.52.55.jpg?:”]2014-05-27 14.52.55.jpg[/url]

[Baite]

Older hosts (non-UEFI) present error

[url=“/_imported_xf_attachments/0/832_2014-05-26 16.00.41.jpg?:”]2014-05-26 16.00.41.jpg[/url]

[Baite]

Just changed undionly.kpxe to be undionly.kpxe.INTEL, and now I managed to have new machines uploading image.
But older machines still getting error in PXE file not found. And they don’t ever start because keep looping in trying to boot from PXE

[Jaymes Driver]

From a windows machine make sure you can even get the file from the server by using this command.

NOTE that if this si windows 7 you need to add the tftp client in the Windows add remove features

[code]
tftp x.x.x.x get undionly.kpxe[/code] where x.x.x.x kis your fog server.

[Baite]

This post is deleted!

[Tom Elliott]

Do they allow for PXE boot? Are they on the same subnet as the systems that are known to work?

[Baite]

From my Windows 7 client, I could download the file (after allowing TFTP in my firewall)

[Baite]

I’m using latest version of FOG (1.1.0) and undionly.kpxe.INTEL (which I renamed to undionly.kpxe).
My server is a CentOS 6.5 installed from zero, with option Basic Server.
Then downloaded Fog 1.1.0 (via SVN) and installed it through ./installfog.sh.
DHCP is handled by CISCO Switch, in a multiple VLAN environement, but all computers boot at the appropriate VLAN (quarentine VLAN, which is where the eth0 NIC card of the server is configured with static IP 10.1.8.1). All configuration (66 and 67) are present and correctly configured (the proof of this is that my 35 new machines all get access to Fog menu, and receive images).
The problem is only with my HP dc7900p computers, which at PXE boot show this error (PXE-E53). They were working perfectly with Fog 0.32 (it’s 350 of them, and all got imaged severall times).
From a Windows 7 client, I managed to get th undionly.kpxe file using “tftp 10.1.8.1 get undionly.kpxe”.

[Baite]

Just tested tftp client frm Windows 7 in one of these HP dc7900p computers, and it worked.
After testing all this, it’s working.
Don’t ask me why or how… but it started working.
Thank you all for your support.