Posts made by Tom Elliott

@Fog_Newb The fos release has been officially updated so if you were to do a full fresh update, it should pull the version you just tested automatically.

Thanks for the testing.

@Fog_Newb Yes, though maybe I can get the master release updated and it take hold for anyone from that point forward who gets 1771, but that’s not great practice, that we fall into better waiting the monthly release I think.

@Fog_Newb Did you update the init to the one you downlaoded after updating to 1771?

The init that it is pulling by default has the bad data.

@Fog_Newb I’m going to close this as it’s a duplicate to the Bug one. Thanks!

@Fog_Newb No.

You shouldn’t have to touch your fog install (overall) to test the latest init files.

Just download the init.xz file from the EXP_20250731 tag and either replace your existing /var/www/fog/service/ipxe/init.xz with it, or put the filename you want to use (init_20250731.xz for example) in the same folder and set the Host’s Host Init field to the filename you created.

Once that’s known to be working, you are welcome to update to the latest version of fog, but you would still have to do the same work for the new experimental init files. We could potentially work to make these working fos files the latest “release” which might fix the issue for you, but first would like to know it’s working.

Luckily init’s are released as a separate funciton so it wasn’t FOGProject 1.5.10.1771 directly, but just the init’s having issues.

@Fog_Newb It seems I introduced a bug on the init’s specifically surrounding a request to fail out if the Host Primary Drive is defined but not found to error out.

Effectively I tried to make a function more robust (normalize output data) but it only works on direct input rather than piped input. I’ve made a correction to fix this miss, but it will take about 2 hours to build.

Maybe I can get you to help?

I am currently building a new set of init’s under experimental: https://github.com/FOGProject/fos/tags

When it completes it will show up as EXP_20250731 (I think?) but if you can download the init and test it on your machine, this should fix the issue you found so thank you for reporting it.

@Fog_Newb For this host in question, are you indicating a serial number or wwn identifier for the drive? On this machine, it seems you have entered 21297930000130911023 on the Host Primary Disk field within the UI. Just trying to understand.

@Greg-Plamondon and @Clebboli are the Storage nodes associated to the same storage group?

For example:

Site 1 -> Default Group
Site 2 -> Default Group

Or are each storage node associated to their own Storage Groups?

The idea of storage groups is to spread the load of imaging. So if all storage nodes are part of the same storage group, then it makes sense the “jumping to different storage nodes within the group”.

While seemingly unexpected, the idea of jumping to other nodes within the same storage group is intended to balance the workload around as best it can. Sure this can probably be better handled, though I don’t know when I might get around to refactoring it. Just saying what I suspect might be happening.

@Clebboii said in Location Plugin - PHP Warning and Location Assigned Node not working:

PHP Warning: Undefined array key “REQUEST_METHOD” in /var/www/html/fog/lib/plugins/location/hooks/addlocationapi.hook.php on line 121

I’ve pushed an update to the dev-branch regarding this issue.

I don’t know if it will definitely fix it, but it should help prevent continued writing of this error.

As for the issue of slot and hopping, I’m not sure how I can simulate this issue. Is this a new issue or something that was noticed and we’re not really sure when it began and/or has always been there?

@shieldsj Can you please provide the exact error from the error logs? What version of FOG are you running?

I suspect the machine you’re attempting to capture the image from does not have an image assigned to it, but without really anything other than there was some error without any real context, not sure what else we can do to help currently.

@SotY The files of console.h/general.h/settings.h were updated 2 - 4 months ago. I don’t think anything here is the cause of the failure. I’ve attempted building and am seeing the same error.

This leads me to think something changed with ipxe code particularly surrounding USB_HCD_USBIO and efi built binaries. It seems the efi builder has a native compiler point for it’s own USBIO that’s being constructed separately from the “common” source.

If you just build the binaries without the USBIO configured, does it break things? I know you found a fix and wanted to try it, but it seems mcb30 (main developer of ipxe) made a change that seems to make the usbio drivers act differently in the past. What those changes were, I do not know at this time.

I am more surprised, in my quick review of things, that this even worked before as the components I see erroring haven’t change significantly.

If I change the usbio_driver line to efi_usbio_driver in the drivers/usb/usbio.c file, and re-run the builder, things build successfully. (Including the ncm–ecm-axge.).

If that works, I can put my diff file and copy it with the build script but I’m scared something I’m doing in the driver/usb/usb.c file might be unintentionally making things worse (not better) even though I’m getting successful builds now.

@sideone No worries. That just let’s us know in your specific case at least, it’s not booting in UEFI mode at least for NIC devices, so thank you!

@sideone THis issue seems completely different from the context of this. I understand, or at least believe the reason this was asked here is because of the “Boot Loop”? portion?

In either case:

If you want to see what I mean, just create an empty file in /tftpboot called autoexec.ipxe (sudo touch /tftpboot/autoexec.ipxe) and the error of “autoexec.ipxe” will disappear. It will not fix the issue you’re seeing, but this message is not, itself, an error. Just the last thing that happens, if I 'm reading the clues correctly.

Do you know what type of NIC the machine that sticks for 5 minutes is using? If it’s realtek or intel, you could try the realtek.efi or intel.efi tftpboot file?

@aurfalien Assuming this is a legacy machine, undionly tries to use the native undi driver of the machine, but you can try other drivers:

ipxe.kpxe might be a good startingpoint and contains more direct drivers for a bunch more devices?

@Small0145 Sure, but it still seems like the golden image had this issue. How it got there I don’t know. I would suggest attempting to boot windows, perform the chkdsk/scandisk as needed. Then shutdown (may need to re-sysprep) after running this command. Windows 11, to my knowledge, can be a little bit strange sometimes.

@Small0145 powercfg /h off

Most of the time, the reason people see this is “shutdown” goes into a special mode called “Fastboot” which works similarly by flipping the dirty bit on the drive.

The command I posted here should prevent “fast log” so when you shutdown, the machine ACTUALLY powers off without any dirty bit.

@skyhawk3355 FOG has been able to be loaded into 256mb of ram for quite some time.

It was originally configured for 128MB of ram, but that was in 2007.

If we know how much RAM is on the actual machine?

I believe all our kernels and init need at least 200mb of ram to operate.

@Eazis YEah, my suspicions correct, this far at least.

You can try downloading rod’s latest kernel:
https://github.com/rluzuriaga/fos/releases/download/EXP_20250617/bzImage

I doubt it will work, but who knows…

@Eazis IF you can boot in and run ip link show

@Eazis Well, what I’m seeing is that the 8125 isn’t being detected at all, and I’m not really sure why at this point. The kernel you tested with has all 8125 firmware associated, as well the driver set to load it. So I’m at a lost as to why it’s not loading that specific thing. The network is detected, but not loaded for some reason. More simply, it seems the kernel can see and tell the type of device, but for whatever reason it doesn’t actually want to load it as a potential interface as being available.

I’m unsure where to go until this portion can work and without one of these myself, it’s really difficult to guess about in the dark.