[quote=“Tom Elliott, post: 43448, member: 7271”]I would LOVE some assistance to figure out this memory issue. I know db size plays some role but it seems a bit ridiculous that we should have to have that much memory to simply read db entries.[/quote]
Is this an issue on the whole web interface or just with snapins or hosts…?

Please make sure you have ‘nfs-common’ (debian/ubuntu see the other posts for package names in centos) installed on [B]the client[/B] from which you want to check the NFS connection!

Portmap is called ‘rpcbind’ in Ubuntu/Debian. Use netstat to see if it is running on [B]the server[/B]:
[CODE]sudo netstat -antup | grep 111
tcp 0 0 0.0.0.0:111 0.0.0.0:* LISTEN 1901/rpcbind
tcp6 0 0 :::111 :::* LISTEN 1901/rpcbind
udp 0 0 0.0.0.0:111 0.0.0.0:* 1901/rpcbind
udp6 0 0 :::111 :::* 1901/rpcbind[/CODE]

I just talked to someone on the iPXE IRC channel and got some helpful information. The RTL8153 should be supported as both the USB NIC and iPXE support ECM! Please try enable more debugging to see what’s going on ‘DEBUG=usb,xhci,usbnet,usbhub’…

And there is another (old) thread here: [url]http://fogproject.org/forum/threads/realtek-8153-usb-network-adapter.10302/[/url]

I suppose you’ve set ‘Legacy only’ in BIOS…

You are absolutely right. Thanks for pointing this out!

It’s being used like that in several locations:
[CODE]src/buildroot/package/fog/scripts/bin/fog.auto.reg:mac=ifconfig | grep HWaddr | head -n1 | base64;
src/buildroot/package/fog/scripts/bin/fog.checkin:mac64=ifconfig | grep HWaddr | head -n1 | base64;
src/buildroot/package/fog/scripts/bin/fog.man.reg:mac=ifconfig | grep HWaddr | head -n1 | base64;
src/buildroot/package/fog/scripts/bin/fog.quickimage:mac=ifconfig | grep HWaddr | head -n1 | cut -d' ' -f 11;
src/buildroot/package/fog/scripts/bin/fog.sysinfo:mac=ifconfig | grep HWaddr | head -n1 ;[/CODE]

I am not sure what would be the best way to clear this up?? Should be register ALL mac addresses or just the connected ones or what?

If I want to make changes to my init files I use this script a lot:
[CODE]#!/bin/sh

INIT=“init”

if [ ! -f ${INIT}.bak ]
then
echo “Backing up ${INIT}.xz”
mv ${INIT}.xz ${INIT}.bak
cp ${INIT}.bak ${INIT}.xz
fi

if [ -f mnt/bin/fog ]
then
umount mnt
rmdir mnt
echo “Compressing ${INIT}.xz, please wait…”
xz -C crc32 -9 ${INIT}
else
xz -d ${INIT}.xz
mkdir -p mnt
mount -o loop ${INIT} mnt
echo “Mounted ${INIT}, have fun…”
fi[/CODE]

Do you see any error messages???

I quite like the idea but I am wondering if it’s worth the effort to create those scripts?! Most problems are very special and some involve switch configurations, DHCP settings on external (windows) servers and possibly other clients are needed to check boot menu via HTTP or NFS mounts.
But feel free to start scripting and we’ll jump in as things get rolling. What do you think?!

You can configure several MAC addresses for one host! Maybe that helps…?

Try adding it to the ‘Parameters’ field in your fog.Deft entry…

Maybe try mounting the NFS share from a different computer to see if this works… Others had broken their NFS shares lately too, see here for example: [url]http://fogproject.org/forum/threads/problem-with-permission-denied-after-fog-server-restart.12503/#post-43257[/url]

Awsome that you got it running now!

Have you talked to iPXE devs too?? Or do you mean FOG devs? I am just wondering because maybe iPXE devs would be able to help out fixing the driver itself. But if this is working for you, you might just leave it like that. Unfortunately I neither have the time nor the hardware to help with this any further although I’d love to help to get this properly fixed.

Seems like there are no images being saved when you upload them. Maybe the are not moved. When you upload an image it’s first saved to /images/dev/<image-name> and moved to /images/<image-name> when it’s finished. Please run this command on the server ‘ls -alR /images’ and post the output.

In your case it would be easier to just rename the files. Please make sure you don’t delete or overwrite any files. I won’t be liable in case anything goes wrong here! Think twice before you copy paste this commands…

[CODE]cd /tftpboot
mv undionly.kpxe undionly.kpxe.orig
mv ipxe.efi undionly.kpxe[/CODE]

Try booting your client. If that works we know that this binary is working for you. We then can go ahead and configure things properly…

Hmmm, that’s interesting. So it’s FOG not being able to save session information properly. Not sure if this is related but someone had issues with wordpress writing zero size session files: [url]http://quirm.net/forum/topic.php?id=3950[/url]
So far I couldn’t really see what was fixing the issue for him. But maybe this can lead us to where we should have a closer look…

From what I could find on the net so far you won’t be able to get iPXE running with this in a rush.

Maybe you are able to use pxelinux without chainloading ipxe.krn but you’ll have to re-think and probably re-write part of the booting process for those machines…

I just tried this:

1. Make pxelinux.0.old the binary that gets loaded by the client after netboot (either by renaming/linking it or by changing DHCP settings).
2. Create a directory /<your-tftp-dir>/pxelinux.cfg and a file called /<your-tftp-dir>/pxelinux.cfg/default:
   [CODE]DEFAULT fog64
   LABEL fog32
   LINUX http://<fog-server-ip>/fog/service/ipxe/bzImage32
   APPEND initrd=http://<fog-server-ip>/fog/service/ipxe/init_32.xz
   LABEL fog64
   LINUX http://<fog-server-ip>/fog/service/ipxe/bzImage
   APPEND initrd=http://<fog-server-ip>/fog/service/ipxe/init.xz[/CODE]
3. Try booting and see if the linux kernel is being loaded! Works for me.
   If it does work for you too you can go ahead and customize this to make if work. You could use DHCP user classes to hand over to pxelinux for this clients and use ipxe for your other clients.

Here are some more options you’ll need to APPEND to the kernel options:
[CODE]root=/dev/ram0 rw ramdisk_size=127000 keymap= web=<fog-server-ip>/fog/ consoleblank=0 loglevel=4 mac=<client-mac-address> ftp=<fog-server-ip> storage=<fog-server-ip>:/images/ storageip=<fog-server-ip> osid=<your-os-id> loglevel=4 consoleblank=0 irqpoll hostname=<client-hostname> chkdsk=0 img=<your-image-name> imgType=<your-image-type_mpa|mps…> imgPartitionType=all imgid=<your-image-id> imgFormat= type=down[/CODE]
Not sure which of those you really need. I just copied them from my testing SVN system. Might be different for FOG 1.2.0…

I don’t think that you’ll be able to integrate this altogether as I can’t think of a way to make pxelinux load the boot.php and all that stuff. But at least you should be able to image those clients I suppose! Please let us know if this is working for you or if you need further hints.

No error message, won’t be easy to find the problem then. Start off with running a “debug upload session”. Click on the host, then Basic Tasks and from there schedule “debug upload”. Hopefully we get more information about what’s wrong with that.

Thank you guys for working on this and making it all run smoothly!

Or maybe ‘secure boot’ related issue, see here: [url]http://fogproject.org/forum/threads/windows-8-1-fog-1-2-0.12528/[/url]

Yes, please upgrade. There have been a lot of bugs and fixes lately…

Maybe this would’ve helped… [url]http://fogproject.org/forum/threads/could-not-boot-exec-format-error-2e008001.12001/[/url]