FOG Project Image Capture on Raspberry Pi 4 (ARM64) via U-Boot

[Jeremy]

@Tom-Elliott

"To answer your question directly: yes, so far I have only been testing manually at the U-Boot prompt (=>).

My main concern right now is scalability. Since I need to deploy this across a fleet of 100 to 200 Raspberry Pi units, configuring each SD card manually or typing commands at the prompt for every single machine is not an option.

What is the recommended best practice in FOG working-1.6 to handle automation for a large fleet? Can the bootcmd be permanently stored in the Pi’s onboard EEPROM once, or can U-Boot fetch a boot script via TFTP automatically without manual card preparation?"

[Jeremy]

@Tom-Elliott

"Hi again! Great news regarding the new TFTP fallback mechanism.

I tested it manually by setting the IP addresses:
Plaintext

setenv ipaddr 192.168.203.50
setenv serverip 192.168.203.113
pxe get

What happened:

U-Boot successfully found the correct path generated by FOG based on the MAC address (Retrieving file: pxelinux.cfg/01-88-a2-9e-53-34-c0).

However, the download times out (Loading: T T T T T T) and results in a Retry count exceeded.

For context: a deployment task was explicitly active and queued in the FOG web interface before running the test, so the file should be there.

Is there a specific TFTP block size or setting required in working-1.6 for U-Boot’s pxe get to successfully pull the file without timing out, or could it be related to firewall/TFTP service configuration on the server side?"

[Tom Elliott]

@Jeremy That’s real progress — pxe get finding pxelinux.cfg/01-88-a2-9e-53-34-c0 by MAC means the naming convention and the file-generation side both work end to end. That’s the part I couldn’t test myself, so thank you for actually running it.

On the timeout itself: “found the file, then hung mid-transfer” (Retrieving file: ... succeeding, then Loading: T T T T T T / Retry count exceeded) points at TFTP’s own connection model, not at FOG. TFTP’s initial request goes to port 69, but the server then hands the rest of that transfer off to a different, randomly chosen UDP port for the actual data — so if there’s a firewall between the Pi and your FOG server that only allows port 69 through, the request succeeds (that’s why U-Boot found the file) and every packet after it gets silently dropped. That matches your symptom closely enough that I’d check it first: either open the OS’s ephemeral UDP range to your TFTP daemon, or (simpler if this is on the same box) confirm nothing’s filtering loopback/LAN traffic to it at all. If you’ve got shell on the FOG server, tcpdump -ni <iface> port 69 or portrange 32768-60999 while you retry pxe get should show the request going out and then nothing coming back, which would confirm it.

Separately — while looking into this I found a real gap on FOG’s side and fixed it regardless of whether it’s your actual cause: the file pxe get downloads wasn’t having its permissions set after upload, so if your TFTP daemon reads as a different user than the account FOG uploads over SFTP with, that alone could produce exactly this symptom (found the file, denied reading it). That fix is in the same working-1.6 branch now, so another update will pick it up. Worth ruling both things out — they’re not mutually exclusive.

On your scalability question: for 100–200 units, don’t script the SD card at all — the RPi 4’s own boot EEPROM supports network boot as the first boot mode, ahead of SD/USB, which is exactly what you want instead of the local-storage-timeout-then-BOOTP-fallback sequence you saw. rpi-eeprom-config (or the raspi-config boot order menu) lets you set that once, and once it’s set it’s a firmware-level property of the board, not something FOG or an SD card image controls — one flash per unit, and after that every power-on goes straight to network boot, no bootcmd persistence step needed on your end at all. I’d flash that as part of whatever imaging/provisioning process gets each Pi ready before it ever talks to FOG the first time, rather than trying to solve it per-deployment.

I don’t have a Pi 4 here to confirm the exact EEPROM boot-order syntax for your firmware revision, so double-check against Raspberry Pi’s own EEPROM config docs before you commit to a fleet-wide flash — but the “make network boot the default at the firmware level” shape of the answer is what actually gets you out of typing anything at the prompt.

[Tom Elliott]

@Jeremy Correction on the firewall angle: if your server went through FOG’s own installer on ufw or firewalld, that’s actually less likely to be it than I made it sound — the installer already handles the exact “TFTP works for the request but the reply is on a random port” problem for both of those (a named tftp service on firewalld, and nf_conntrack_tftp loaded for ufw), specifically because it’s bitten people before.

So rather than “check your firewall” in general, four narrower things:

1. Which firewall are you actually running — ufw, firewalld, or plain iptables? The installer only auto-configures the first two; if it’s bare iptables, my original theory stands and you’d need to allow the ephemeral UDP range by hand.
2. If it’s ufw: lsmod | grep tftp — the installer tries to load nf_conntrack_tftp but silently continues if that fails, so it’s worth confirming it’s actually loaded rather than assumed.
3. Is BOOT_external_tftp_server set to “yes” in your FOG settings? If so and your TFTP is actually running on the FOG box itself, firewalld would have skipped opening it for you.
4. Are the Pi and the FOG server on the same subnet/VLAN, or is there a router, switch ACL, or (if this is a VM/cloud box) a security group between them? None of the above touches anything outside this box.

If you’ve got shell on the FOG server, tcpdump -ni <iface> port 69 or portrange 32768-60999 while you retry pxe get is still the fastest way to see whether the request goes out and nothing comes back (firewall/network) versus the request itself failing (something else entirely).

[Jeremy]

@Tom-Elliott

Here is an update following our latest troubleshooting sessions and tests:

The !V@JL artifact: That part was indeed a leftover artifact from editing the file manually with nano. Once cleaned up, we hit a regression: the Raspberry Pi no longer downloads the 563-byte config file at all anymore. It now loops directly on BOOTP broadcast and times out without hitting the TFTP stage (Retrieving file... doesn't trigger anymore).

Goal & Architecture: The target deployment must remain 100% cardless (no local storage, no SD card, no USB drive used during boot), relying purely on the RPi 4 network boot EEPROM + OPNsense DHCP + FOG/U-Boot.

Regarding your earlier notes on TFTP/Firewall:

    We are running UFW on the FOG server, and we verified that nf_conntrack_tftp is active (lsmod | grep tftp confirmed).

    BOOT_external_tftp_server is not enabled (TFTP runs locally on the FOG box).

    The Pi and FOG server are on the same subnet, with no intermediate restrictive ACLs on that local segment.

The permission fix: We noted your mention about the file permissions fix for downloaded files not being set properly after upload in the working-1.6 branch. Could this regression (dropping back to infinite BOOTP loops instead of attempting the TFTP fetch) be tied to how the file is currently served or indexed, or is there a specific way U-Boot expects the handoff after DHCP?

[Tom Elliott]

@Jeremy Good — that rules out the firewall/permissions angle cleanly, thanks for actually checking rather than taking my word for it.

The new symptom is a different failure than before, and it’s outside anything FOG serves: looping on BOOTP with no attempt at Retrieving file... means it’s failing at DHCP/BOOTP negotiation, a step before pxe get would ever run. FOG can’t see that far back — the board never gets far enough to ask FOG anything.

You’ve also moved to a materially different boot path than what we’d been testing: cardless, via the Pi’s own network-boot EEPROM, rather than typing commands by hand at the U-Boot prompt. That matters here, because the EEPROM’s automatic network-boot flow runs its own boot sequence, built into that U-Boot, not necessarily the dhcp / pxe get / pxe boot lines from earlier in this thread — those were for a manually-typed bootcmd. If the EEPROM path uses bootp instead of dhcp, or expects OPNsense to hand it specific DHCP options, that’s a different thing to get right than what we tested manually.

Two things I’d need to actually say anything useful here, since I can’t see either from where I’m sitting:

1. What’s the board actually running right now — did the nano edit touch a persisted bootcmd (printenv bootcmd), or is this the EEPROM’s own default network-boot sequence with nothing custom in the loop at all?
2. In OPNsense’s DHCP config for that subnet, what are options 66 (next-server) and 67 (filename / bootfile-name) set to? The EEPROM’s PXE client needs those to know where to send its own request in the first place — if they’re pointing at the wrong place, or missing, U-Boot never gets the chance to see FOG at all, and that would produce exactly a BOOTP loop with nothing after it.

That’ll tell us whether this is a boot-script problem (fixable on your end) or a DHCP-options problem (fixable in OPNsense) — right now I genuinely can’t tell which from here.

[Jeremy]

@Tom-Elliott

Hi, I’m sending you the two screenshots—one from OPNsense and the other from the Raspberry Pi.

[Tom Elliott]

@Jeremy Thanks for the screenshots — they changed my read of this quite a bit, and I owe you a correction first.

Correction: earlier I said pxe get “finding” pxelinux.cfg/01-88-a2-9e-53-34-c0 proved the file-generation side worked. It didn’t. U-Boot prints Retrieving file: ... before it sends the request, and the Loading: T T T T that follows means the server never answered at all. A file that really is missing gives you TFTP error: 'File not found', not timeouts. So the file-side was never proven, and it turns out it was broken on my side of the fence.

The FOG bug, now fixed: FOG was writing the 01-<mac> file under the directory it keeps the HTTP-served kernels in (/var/www/html/fog/service/ipxe/), not under /tftpboot. The TFTP daemon runs chrooted to /tftpboot, so the file existed but TFTP could never see it. That’s fixed in working-1.6 (PR #1664) with a new setting, FOG Settings → TFTP Server → FOG_TFTP_ROOT_DIR, which the installer sets to the real TFTP root. Run the updater once it’s merged and check that setting reads /tftpboot; after that, queue a task for the Pi and you should see /tftpboot/pxelinux.cfg/01-88-a2-9e-53-34-c0 appear on the server. Please confirm that file is there before the next boot test — it’s the one thing I can check in code but not on your box.

Your current boot loop is a different, earlier step. From the console shot: U-Boot itself loaded cardless, which means the Pi’s EEPROM network boot already pulled the firmware and u-boot.bin from /tftpboot over TFTP. That proves the OPNsense next-server, the FOG TFTP service and the firewall are all fine for this client — so ignore my earlier firewall angle. What’s looping is bootcmd=bootflow scan: that’s U-Boot’s standard-boot sequence, which tries mmc, usb, then ethernet, and the ethernet step starts with U-Boot’s own DHCP request (BOOTP broadcast 1, 2, 3...). That DHCP never succeeds, and it’s never been shown to work in this thread — your earlier manual test set ipaddr/serverip by hand and skipped DHCP entirely. The empty per-architecture filenames in OPNsense are fine, by the way: the pxe boot method doesn’t need a bootfile name at all. (Side note, unrelated to the Pi: the iPXE-class filename in OPNsense is what gets handed to a client that is already running iPXE, and FOG expects default.ipxe there, not ipxe.efi — with ipxe.efi your x86 UEFI clients will reload iPXE forever. Worth a look when you’re back on those.)

Two things to try at the U-Boot prompt, and please paste the exact text rather than a screenshot if you can:

dhcp

on its own. If it also loops on BOOTP, then U-Boot’s network driver isn’t getting a lease and we look at OPNsense’s DHCP log for that MAC (does the request even arrive, does it offer). If it does get an address, then:

pxe get
pxe boot

once the FOG update above is in and the 01-... file is confirmed on disk. If that works by hand, setenv bootcmd 'dhcp; pxe get; pxe boot' and saveenv gets you the automated path — bootflow scan should get there too, but the manual sequence tells us which step is at fault when it doesn’t.

[Jeremy]

@Tom-Elliott

I do have the filename you requested.

When U-Boot starts, DHCP doesn’t work; it spends 8 seconds searching. Sometimes, when rebooting the Raspberry Pi, it manages to obtain an IP address, successfully detects the server’s IP, and downloads pxelinux, but then stops.

What I don’t understand is why U-Boot sometimes manages to get an IP address and sometimes doesn’t.

[Tom Elliott]

@Jeremy Do you happen to have STP on your network?

[Tom Elliott]

@Jeremy Good — the file being there confirms the /tftpboot fix landed. Two separate things in your report, and one of them was mine again.

“Downloads pxelinux, but then stops” — partly expected, partly another FOG bug, now fixed.

• If there was no task queued for the Pi at that moment, the file FOG writes says localboot 0, which tells U-Boot “nothing to do, carry on to local disk”. That’s by design: no task, no imaging. So “stops” after a successful pxe get is correct behavior unless you’d queued a capture first.
• If there was a task queued, it would still have stopped, and that one is on me: the file named the kernel and init as http://… URLs, which is what boards with wget use. Your U-Boot’s pxe code can’t follow a URL — I checked U-Boot’s source (boot/pxe_utils.c every kernel and initrd line is fetched over TFTP, relative to wherever the config came from, so kernel http://… became a TFTP request for a file literally called that. Fixed in working-1.6 (PR #1679): the file now says kernel arm_Image / initrd arm_init.cpio.gz, and FOG copies those two files into /tftpboot itself the first time a task is queued for an ARM host, and again whenever the kernel is updated. Run the updater once more, queue a capture for the Pi, then on the server check that /tftpboot/pxelinux.cfg/01-88-a2-9e-53-34-c0 contains kernel arm_Image and that /tftpboot/arm_Image and /tftpboot/arm_init.cpio.gz exist.

The intermittent DHCP is a network-timing problem, not a FOG one, and Tom’s STP question is the right first suspect. On a switch port running classic spanning tree, the port doesn’t forward traffic for roughly 30 seconds after link-up (listening, then learning). U-Boot brings the link up and sends its DHCP discover immediately, retries for a few seconds, and gives up — that’s the “8 seconds searching”. The times it works are the times the port happened to already be forwarding. The Pi’s EEPROM boot succeeds because it retries for much longer. Two fixes, and I’d do both:

• On the switch: enable PortFast / edge-port (or RSTP) on the ports the Pis plug into. That’s the real fix and it helps every PXE client, not just the Pis.
• In U-Boot, so a slow port doesn’t kill the boot anyway:

setenv autoload no
setenv netretry yes
setenv bootcmd 'dhcp; pxe get; pxe boot'
saveenv

autoload no stops dhcp from also trying to TFTP a bootfile it was never given (that’s an extra timeout and a spurious failure in your sequence). netretry yes makes U-Boot keep retrying DHCP instead of giving up after a few seconds — the right call for a headless board that has nothing else to do, but be aware it means a Pi with no DHCP server on the wire will sit there retrying rather than dropping to the prompt.

If DHCP still fails after that, the next read is OPNsense’s DHCP log for that MAC: whether the discover even arrives, and what it offers. And as before, pasted text beats a screenshot — the exact lines around pxe boot are the ones I need next.