FTP protocol problem

December 24, 2016, 5:16 am

≫ Next: http authentication bug after daa8 commit?

Hi everyone ,

I did buy last week a router from assus ( RT-N18U) . Router have gigabit lan and 2 USB ports ( one usb 3.0 port and one usb 2.0 port) . So I conected a usb 3.0 stick to router and configured FTP server so i can acces those files from stick via FTP . I took latest ipxe source from github and recompiled to include FTP suport .I did also embeded a script that contain 2 lines :

dhcp
chain ftp://user:password@192.168.2.1/sda/ipxe/menu.ipxe

Then I uploaded the undionly.kpxe to router and configured dnsmask.conf.
FTP server on the router is "" vsftpd - 2.3.4-1 "" .

Now when I started the client to boot this is what happend :

1 . Client did request a IP and boot file from dnsmask .
2. Dnsmask received the request and alocated a dynamic IP to client and give the boot file name .
3. client downloaded the undionly.kpxe file from TFTP and released control to undionly .kpxe
4. undionly.kpxe detected the network card , initialized and then launched the commands from the EMBEDed script .
5 .Client requested IP again from dhcp and then downloaded the menu.ipxe from the ftp server.
6.Menu launched succesfully and downloaded ipxe.png from same ftp server.
7.Then I launched tinycore linux that use next code in the menu.ipxe :

:tinycore
echo Booting Tinycore (With Gui)
set url ftp://user:password@192.168.2.1/sda/ipxe...es/liveiso
initrd ${url}/tinycore/tinycore.iso
chain ${url}/tinycore/memdisk iso || goto failed
goto start

When the ISO image need to be downloaded from ftp server somthing very strange happend . The nr at the end of line that indicate how much % of the file is downloaded in real time go in about 0.001 seconds from 0% to 100% and then everything freez on screen .The file is never downloaded and other lines never executed as a result of this behavior . If i press CTRL+C to abort the download that work and i got an error that is no OS to load and i need to reboot . In order to figure out if ipxe is doing something or is completly frozen , I did started another ftp client and start download something from same ftp server .I noticed that download speed droped from 50 mb/sec to 25% mb/sec on the second download. Its like ipxe is practically using the remaining 25 mb/sec , trying desperatly to download that ISO file but never succeding. I did tested like 15 diff FTP clients with the ftp server and all 15 working fine except ipxe .
I did atached a print screen with the problem below .

↧

http authentication bug after daa8 commit?

December 28, 2016, 4:24 am

≫ Next: Can't boot Debian Jessie using preseed (UEFI)

≪ Previous: FTP protocol problem

I'm trying to use HTTP Digest authentication using login example from http://ipxe.org/cmd/login, using the standard iso build

Code:

login

chain http://${username:uristring}:${password:uristring}@my.web.server/boot.ipxe

But the chain command returns with "Permission denied (http://ipxe.org/020c613c)

I thought that it could be a problem with the web server, but when I try the same URL in a browser or using 'wget' (with user and password options) it works without problems.

I used tshark to review the network, and I can see that:
- The iPXE client sends an HTTP get, and the server answers with a 401 reply
- The iPXE client then sends another HTTP with the user and password information, and the server replies with a 200 OK and the contents of the file

But even with the server sending the correct information, the iPXE client shows the "Permission denied" error.

I've tested this problem with Basic and Digest authentication.

Looking for information in the forums I've found that the HTTP Authentication was working, but that it has failed lately to at least another user (http://lists.ipxe.org/pipermail/ipxe-dev...5263.html) with the same symptoms (401 followed by 200 and still an error).

I've started to test older iso.pxe builds (downloaded from rom-o-matic), using Ctrl-b to enter the iPXE shell and then running:

Code:

dhcp

initrd http://user:hello@my.server/protected/file.ipxe

And I've found that with the commit

- https://git.ipxe.org/ipxe.git/commit/b99...d1e63230c9 (Disable TIVOLI_VMM_WORKAROUND in the qemu configuration) HTTP authentication worked OK
- but with the next build https://git.ipxe.org/ipxe.git/commit/daa...2c98b0a11a (Provide intf_reinit() to reinitialise nullified interfaces) the HTTP Authentication didn't work.

Please, could someone test this behaviour?

I reverted some changes in the httpcore.c and interface.c from the daa8 commit. Then recompiled the iso.pxe and it worked. It seems to be a problem with the http->content not being cleaned between the 401 and the 200 responses, but I really don't have the expertise with the code to provide a patch.

Will continue to use the b991 code by now.

And thanks for this fantastic project: it's incredible what you can achieve with it, and is been really a pleasure to work with.

↧

Can't boot Debian Jessie using preseed (UEFI)

December 28, 2016, 6:03 am

≫ Next: Cannot flash Intel nic: "invalid image file"

≪ Previous: http authentication bug after daa8 commit?

Hey, guys!

I've been trying to boot a Debian system using this guide.

Unfortunately, the preseed.cfg file is not added to the initrd, I can't see it on a debian-installer filesystem, and debian-installer does not use preseed (obviously, since it cannot find it).

I am using the latest ipxe.efi executable.

Here is my iPXE script:

Code:

#!ipxe

kernel linux initrd=initrd.gz initrd=preseed.cfg

initrd initrd.gz

initrd preseed.cfg preseed.cfg

boot

I'd be really glad for any help I can get, I've been banging my head against the wall entire day.

↧

Cannot flash Intel nic: "invalid image file"

December 29, 2016, 9:52 am

≫ Next: rom-o-matic PXE_STACK debug build

≪ Previous: Can't boot Debian Jessie using preseed (UEFI)

I'm trying to flash an intel nic with iPXE. I have been following the docs and got my nic ID (80861502) made the rom, downloaded bootutil, booted freedos but the -restorimage command fails with the error "Invalid image file"
I found on the forums to check the ROM image size and I can see the rom created from make is 65K which appears correct and I also saw recommendations to try older versions of bootutil. I tried versions 18.3, 19, 20, and 21 and all give the same error. I haven't tried flashrom but am begining to think there may actually be something wrong with my rom.

How can I verify I have a good image and does anyone have suggestions for where to look next to get the rom flashed?

↧

rom-o-matic PXE_STACK debug build

January 4, 2017, 10:24 pm

≫ Next: iPXE EFI build fails with elf2efi64 failed

≪ Previous: Cannot flash Intel nic: "invalid image file"

I decided to give https://rom-o-matic.eu/ a go today, I did not know had been updated, last time I checked it was still gpxe, great! I was able to build some .efi test files which work great, and is very convenient than compiling it manually with changing options in vim or nano etc

It appears that ticking the PXE_STACK option with EFI PXE bootstrap 64-bit .efi results in a debug crash everytime, however I don't know if this option is necessary since it seemed to keep my IP address when loading ipxe anyway.

Not sure if this an issue but anyway thanks for the great work with updating rom-o-matic to ipxe.

↧

iPXE EFI build fails with elf2efi64 failed

January 6, 2017, 12:06 am

≫ Next: iPXE menu

≪ Previous: rom-o-matic PXE_STACK debug build

When I try to build the ipxe.efi with latest version of ipxe

The command I am using is-----

Code:

make bin-x86_64-efi/ipxe.efi

it comes with an error-----

Code:

[HOSTCC] util/elf2efi64

/tmp/cc5uKw9w.o: In function `elf2pe':

/home/user/Desktop/ipxe/src/util/elf2efi.c:933: undefined reference to `__xpg_basename'

/usr/bin/ld: util/elf2efi64: hidden symbol `__xpg_basename' isn't defined

/usr/bin/ld: final link failed: Bad value

collect2: error: ld returned 1 exit status

Makefile.housekeeping:1375: recipe for target 'util/elf2efi64' failed

make: *** [util/elf2efi64] Error 1

↧

iPXE menu

January 4, 2017, 8:48 pm

≫ Next: Running console command with --keep option -> unable to boot

≪ Previous: iPXE EFI build fails with elf2efi64 failed

Newbie here.. still trying to understand the iPXE setup and usage scenarios (vs PXE).

With PXE, I have a loader (pxelinux.0) and a menu (/pxelinux.cfg/default). The menu is a text file, sits on the TFTP server, and is easily editable on the fly.

With iPXE, I have a loader (undionly.kpxe), however I understand any further instructions have to be embedded into the loader via make command.. which is a pain (Windows only environment here).

Is there a way to make iPXE loader to load some default *.ipxe menu file from the same TFTP server ? Just trying to find a way to easily edit a text file menu (as in PXE case), without having to customize the undionly.kpxe every time..

↧

Running console command with --keep option -> unable to boot

January 7, 2017, 3:54 am

≫ Next: Sanboot timeout 4cC8520 with Apache

≪ Previous: iPXE menu

Hello, I'm fairly new to iPXE but I hope I can supply you with all the relevant info.

I'm unable to boot any image/kernel I tried if I use the console command with --keep option.
iPXE version used: 1.0.0+ (26050)
I tried both the ipxe.kpxe and undionly.kpxe files.

Tested with:
VMware Workstation 12 Player version 12.5.2 build-4638234 (32-bit VM)
and on real hardware (64-bit) (can post specs if needed).

I'm using dnsmasq on my router (DHCP) and my boot server (Proxy DHCP).

Image used:
Memdisk and memtest from http://ipxe.org/appnote/memtest

When I run console with --keep option before booting an image I get the following error message:

Code:

El Torito BVD sanity check failed.

El Torito boot catalog sanity check failed.

MEMDISK: Image seems to have fractional end cylinder

MEMDISK: Image appears to be truncated

After I press any key to boot, the machine (vm or real hardware) just locks up and I have to hard reset it.

This led me to believe that there was some corruption happening somewhere so I ran the tests at http://ipxe.org/dev/drvtest/md5sum
But to my confusion it passed the tests flawlessly.

iPXE compiled with:

Code:

make bin/{undionly,ipxe}.{,k,kk}pxe EMBED=../boot.ipxe

My scripts:
Embedded boot.ipxe

Code:

#!ipxe

:findserver

echo Trying DHCP... ||

dhcp || goto nodhcp

isset ${proxydhcp/dhcp-server} && set next-server ${proxydhcp/dhcp-server} ||

isset ${proxydhcp/next-server} && set next-server ${proxydhcp/next-server} ||

isset ${next-server} || goto noserver

echo Boot server found: ${next-server} ||

:chain

chain --replace --autofree http://${next-server}/ipxe/menu.php?platform=${platform}&mac=${mac} ||

set errmsg ERROR: Chainloading failed ||

prompt Press any key to continue... ||

goto error

:nodhcp

set errmsg ERROR: Cannot reach DHCP server ||

prompt Press any key to continue... ||

goto error

:noserver

set errmsg ERROR: Cannot locate boot server ||

goto error

:readip

echo -n Please enter the boot server ip address: && read next-server && goto chain ||

:error

menu -- ${errmsg} ||

item --key r -- findserver (r) Retry ||

item --key e -- readip (e) Enter ip manually ||

item --key c -- config (c) View config ||

item --key s -- shell (s) Open shell ||

item --key b -- reboot (b) Reboot ||

item --key n -- shutdown (n) Shutdown ||

item --key x -- exit (x) Exit ||

choose --default ${action} action || goto exit

goto ${action}

:config

config ||

goto error

:shell

shell ||

goto error

:reboot

reboot ||

goto exit

:shutdown

shutdown ||

:exit

echo Booting from local hard disk... ||

menu.php (I removed the php code and menu for debugging)

Code:

#!ipxe

console --x 800 --y 600 --picture menu.png --keep --left 80 --right 80 --top 60 --bottom 60

initrd files/images/memtest.iso

chain files/images/memdisk iso raw pause

I also tried booting into Windows 7 setup (from http://ipxe.org/howto/winpe)
menu.php

Code:

#!ipxe

console --x 800 --y 600 --picture menu.png --keep --left 80 --right 80 --top 60 --bottom 60

kernel files/windows/wimboot

initrd files/windows/boot/bcd BCD

initrd files/windows/boot/boot.sdi boot.sdi

initrd files/windows/sources/boot.wim boot.wim

boot

It fails with the following error message:

Code:

Bad CPIO magic

FATAL: could not extract initrd files

Press a key to reboot...

And finally I tried gparted live:
menu.php

Code:

#!ipxe

console --x 800 --y 600 --picture menu.png --keep --left 80 --right 80 --top 60 --bottom 60

kernel files/gparted/i386/vmlinuz boot=live config components union=overlay username=user noswap noeject ip= vga=788 fetch=http://192.168.1.2/ipxe/files/gparted/i386/filesystem.squashfs quiet splash

initrd files/gparted/i386/initrd.img

boot

This failed with:

Code:

Initramfs unpacking failed: junk in compressed archive

Everything boots just fine without the --keep option.
Am I missing something here? What am I doing wrong?
I would include screenshots too but attachments are disabled I guess.
Feel free to ask anything, I can post more configs if needed.

↧

Sanboot timeout 4cC8520 with Apache

January 7, 2017, 6:00 pm

≫ Next: WDS, iPXE, DHCP and SecureBoot

≪ Previous: Running console command with --keep option -> unable to boot

Hello Everybody,

I'm newbie of ipxe, pls help me for this issue:

I can sanboot with iso file in IIS 7.5 well but timeout error 4cC8520 with same that iso file over Apache24 (Apache and IIS hosted in same computer)
Although that iso file small, also test same error with other kind of file.
Anotherway, i still can http boot with that Apache server by other iPXE way as initrd =>boot for winpe.wim example.

When check Apache access LOG, it show HEAD command only, not GET command from client as IIS LOG

Check with Wireshark (from Apache/IIS computer) when have that error, there is one error:
===========
Acknowledgement number: Broken TCP. The acknowledge field is nonzero while the ACK flag is not set
============

So I suspect that ipxe tcp stack have issue with Apache, or i need adjust something in Apache configuration.

Any idea Pls help!
Thank You in advance!

↧

WDS, iPXE, DHCP and SecureBoot

January 11, 2017, 7:24 am

≫ Next: iPXE and USB-C

≪ Previous: Sanboot timeout 4cC8520 with Apache

Hello,

Currently I am using iPXE as a menu on all devices except SecureBoot enabled devices. For SecureBoot enabled devices, I'd like to always boot to WDS;

I suspect that DHCP has to provide this to the client, based on the fact if it's secureboot enabled, or not.

Since iPXE has a lot of DHCP options, i was wondering if this was possible; is DHCP aware of secure boot? And if so, what DHCP class option is this?

btw, I am using Linux DHCPd.

On another server, i chain the other way around (so the devices boot to WDS, and WDS provides those devices with PXElinux (configured via:

wdsutil /set-server /bootprogram:boot\x64\pxelinux.com /architecture:x64
wdsutil /set-server /N12bootprogram:boot\x64\pxelinux.com /architecture:x64
)

Is that same situation available for iPXE?

↧

iPXE and USB-C

January 12, 2017, 3:58 am

≫ Next: iMAC with Broadcom 14e4-1686 BCM57766 : link status down

≪ Previous: WDS, iPXE, DHCP and SecureBoot

Hi list

I am using iPXE to PXE-boot Macs to a WDS server in order to install Windows on them. For older Macs (with BIOS-emulation), I use the prebuilt ipxe.iso-Image, for the newer ones (the ones with EFI-boot) I use an ipxe.efi executable on a USB stick. This has worked fine so far, even for Mac laptops that do not have a built-in NIC and where I have to use a Thunderbolt-ethernet or USB-ethernet adapter.

However, the new MacBooks Pro now only have combined USB-C/Thunderbolt 3 ports, and no matter what kind/brand of network adapter I use, I am not able to PXE boot them. Either the network adapter is not seen at all, or it is seen as down, or it is seen and up, but the configuration fails.

Unfortunately, I am not really an expert when it comes to PXE-booting... I suspect that maybe a driver for USB-C is missing? Any ideas what I could try? This is really a show-stopper for us.

Thanks, Tina

↧

iMAC with Broadcom 14e4-1686 BCM57766 : link status down

January 18, 2017, 8:06 am

≫ Next: Chain grub/grub2 failed

≪ Previous: iPXE and USB-C

Hi

Working with FOG, we have an iPXE issue with iMac Late 2015 with a Broadcom ethernet adapter.

Here is the thread on FOG forum :
https://forums.fogproject.org/topic/9269...-ipxe-boot

To resume it, when we boot on ipxe :

Code:

IPXE initialising devices...ok

Waiting for link-up on net0... failed: Down (http://ipxe.org/38086193)

On ipxe shell, ifstat say :

Code:

net0: 38:c9:86:xx:xx:xx using 14e4-1686 on 0000:03:00.0 (closed)

  [Link: down, TX:0 TXE:0 RX:0 RXE:0]

  [Link status: Down (http://ipxe.org/38086193)]

In the ipxe shell, we have tried to set up ip manually.

We have made our own ipxe.efi with https://rom-o-matic.eu with :
- all drivers
- only 14e4-1686

We have tried also all the efi image provided by FOG developers.

We have changed RJ 45 cable.
We have portfast activated on the switch port and tested with a non manageable switch. The link LED is always on.

Is there a problem with the driver ?
Thanks for help.

↧

Chain grub/grub2 failed

January 20, 2017, 5:16 am

≫ Next: How to get UEFI booting from USB stick to work?

≪ Previous: iMAC with Broadcom 14e4-1686 BCM57766 : link status down

Hey, hope this is not a beginner mistake.

I want to chain grub and/or grub2 for installing CentOS6 and/or CentOS7 because trying to load kernel and initrd directly from iPXE failed.

I have a script for iPXE which looks like this:

Code:

#!ipxe

chain tftp://xxx.xxx.xx.xx/centOS7/grubx64.efi

I always end up in the grub shell. My first thought was that the grub.cfg file is not loaded but I don't know how to.

Thanks for your help.

Additional Information:

Handling grubx64.efi directly from my dhcpd.conf works fine for me. So this:

Code:

class "pxeclients" {

             match if substring (option vendor-class-identifier, 0, 9) = "PXEClient";

              next-server xxx.xxx.xx.xx;

             if option architecture-type = 00:07 {

                    filename "centOS7/grubx64.efi";

              } else {

                    filename "xxxx";

              }

in dhcpd.conf works fine, but

Code:

filename "ipxe.efi";

with the script above doesn't work and ends up in the grub shell.

↧

How to get UEFI booting from USB stick to work?

January 25, 2017, 7:29 am

≫ Next: Can't figure out to to boot to my SCCM Server

≪ Previous: Chain grub/grub2 failed

Hello,

I have been trying to install iPXE on USB stick, so I could start PXE install from Clonezilla server.

But I don't get it working. When I try to write the ISO file taken from the website using Rufus to the USB stick, Rufus tells that the ISO file is not UEFI compatible.

Also, when I have tried to format the memory stick to FAT32 and manually try to start ipxe.efi from shell, I get error message "Could not boot image: Exec format error"

Are there any step By step tutorials how to make UEFI PXE booting working? The BIOS boot option is not available on the Computers, only UEFI.

↧

Can't figure out to to boot to my SCCM Server

January 25, 2017, 8:11 am

≫ Next: UEFI, iPXE, and HP Service Pack for ProLiant (SPP)

≪ Previous: How to get UEFI booting from USB stick to work?

I have 2 VMs on the same machine. One is a System Center Configuration Manager 2016 server that is configured to allow PXE requests and has a boot image ready to go on the distribution point. It also does DHCP and DNS, along with AD. Its FQDN is SCPROTO.tst

The second is just a test desktop VM with no OS installed. I boot directly to iPXE and can get an IP from the VM server. But I can't seem to figure out what command to use to get it to use pxe to boot to it.

Any suggestions?

↧

UEFI, iPXE, and HP Service Pack for ProLiant (SPP)

January 25, 2017, 12:36 pm

≫ Next: Flash iPXE to PCI NIC SysKonnect SK-9521 V2.0

≪ Previous: Can't figure out to to boot to my SCCM Server

I am trying to deploy SPP over the network using EFI, it works when in legacy mode, but soon we will be switching over to EFI. The servers are ML110s, ML350s, DL380s, and more...

The current message that keeps being displayed is:

Code:

Kernel panic - not syncing: VFS: Unable to mount root fs on unknown-block(0,0)

CPU: 0 PID: 1  Comm: swapper/0 Not tainted 3.12.38-44-default #1

SPP is 2016-10

iPXE is compiled from git source on 2017-01-17 build 26050

I also tried chainloading:
syslinux.efi, it crashes.
elilo reports token is too long.
grub2 is unable to get an IP address, and using a static IP is not a reasonable option because more than one server will be running SPP and there is more than one subnet.

the ipxe menu contains:

Code:

kernel http://${ipserver}/data/hp_spp/${hpsppver}/vmlinuz initrd=initrd.img rootfstype=ramfs

initrd http://${ipserver}/data/hp_spp/${hpsppver}/initrd.img

imgargs vmlinuz media=net rw ramdisk_size=1099688 root=/dev/ram0 init=/bin/init iso1mnt=/mnt/bootdevice loglevel=3 ide=nodma ide=noraid nopat pnpbios=off vga=791 splash=silent hp_fibre showopts noexec32=off numa=off nox2apic ipappend 2 TYPE=MANUAL AUTOPOWEROFFONSUCCESS=no AUTOREBOOTONSUCCESS=no iso1=nfs://${ipserver}/opt/planb/data/hp_spp/${hpsppver}/${hpsppiso} iso1opts=timeo=120,nolock,bg,ro

boot

↧

Flash iPXE to PCI NIC SysKonnect SK-9521 V2.0

January 31, 2017, 8:57 am

≫ Next: Having issues configuring DHCP

≪ Previous: UEFI, iPXE, and HP Service Pack for ProLiant (SPP)

I have Gigabit NIC SysKonnect SK-9521 V2.0 with SPI flash. I make dump 128KB flash with yukondg.exe DOS utility. Using this tool can also probably flash firmware - Load Boot Code from File:
[Image: 8851214.png]

Is this tool can safely flash iPXE? Will I be able to go back to the original firmware?

↧

Having issues configuring DHCP

January 30, 2017, 11:40 pm

≫ Next: Computer reboots after loading boot.wim

≪ Previous: Flash iPXE to PCI NIC SysKonnect SK-9521 V2.0

I am trying to chainload iPXE from PXE so that i can pull an image from the web and boot from it. I set up a Windows 2012 server and followed the instructions on http://ipxe.org/howto/chainloading and http://ipxe.org/howto/msdhcp#pxe_chainloading however, it doesnt seem to be working. When booting, i get the following error:
PXE-T04: Access Violation
PXE-E36: Error received from TFTP Server.
PXE-M0F: Exiting Intel Boot Agent.
[Image: DtKWmlAU1ZIz9KZec6pflEStif3Cx78hMj16JFsV...ize_mode=3]

[Image: DtKWmlAU1ZIz9KZec6pflEStif3Cx78hMj16JFsV...ize_mode=3]

I dont even get to the iPXE banner.

My DHCP config is below and also attached
[Image: jCmq_91zLf4oBG9BGRkd8LtczKYN8bUhru_K_mIz...ize_mode=3]

[Image: Wx6ghATrX_v0gkxkFNy0umW3DVH2lO6q7zOpYSo_...ize_mode=3]

Ive been working on this for a couple of days now and any help would really be appreciated

Thanks!
Mony

↧

Computer reboots after loading boot.wim

February 2, 2017, 6:49 pm

≫ Next: Boot to ipxe on UEFI machine

≪ Previous: Having issues configuring DHCP

I am chainloading iPXE then using wimboot to boot to WinPE. However, when booting up, the computer reboots after loading boot.wim. Has anyone experienced this before?

The code for boot.ipxe is the following:

Code:

#!ipxe

  kernel wimboot

  initrd boot/bcd         BCD

  initrd boot/boot.sdi    boot.sdi

  initrd sources/boot.wim boot.wim

boot

A link to a snapshot of the boot sequence is below
https://www.dropbox.com/s/aeh0pcn9yvfb2s...0.PNG?dl=0

It reboots right after this.

↧

Boot to ipxe on UEFI machine

February 3, 2017, 4:43 am

≫ Next: iPXE Not Making Requests to Port 4011

≪ Previous: Computer reboots after loading boot.wim

I have a simple setup of

server1= SCCM 2012 R2
server2= MS DHCP server

MS DHCP option 66/67 are configured for ipxe clients (legacy) to boot to undionly.kkpxe created with embedded script that does:

#!ipxe
dhcp net0
set 209:string pxelinux.cfg/default
set 210:string http://sccm.domain.com/tftpboot/
chain ${210:string}pxelinux.0

That works perfect for legacy PXE boot

Tried to follow DHCP policies as per:

https://wiki.fogproject.org/wiki/index.p...HCP_Policy

to get it boot to efi.0 (made with same script) from EFI network boot on physical hardware

But none of my attempts was ever successful
each time such machine boots directly only to \Boot\x64\bootmgfw.efi from SCCM server

Anybody any ideas how/what to change to make it boot to the menu?

sebus

↧

Latest Images