An MCL import merged Deploy/Applications with rsync -a, which replaced each
image's BPRT symlink with a plain copy of the vendor folder. That copy has
no site overlay, so standard/engineer enrollment lost
Register-BulkEnrollTask-Standard.ps1 (specialize Order 39).
Now Deploy/Applications is imported first and without BPRT:
- where _shared/Applications/BPRT exists it is left alone (the BPRT page
keeps it current) and the image link is put back;
- on a box with none yet, the imported BPRT seeds it: vendor
Select-PCConfig.ps1 kept as -vendor, our overlay added, BOM fix applied;
- the import refuses before anything moves when that folder would not work
(vendor scripts missing, or no overlay copy to add).
The two Register scripts join Select-PCConfig.ps1 in the repo. The playbook
already syncs shopfloor-setup to the enrollment share, so
config.BPRT_OVERLAY_DIR finds all three there; overlay_sources now falls
back to it for every overlay file, not only the shim.
Engineer builds left the employee without "Configure my Engineering
Layer". First logon ran C:\Deploy\Applications\extra_engineering\
Install-EngineeringLayer.ps1, and the script itself lived in C:\Deploy:
std/eng have no autologon, so a PC that rebooted before the employee
signed in had lost C:\Deploy to the orchestrator's DeleteDeployDir task,
and nothing installed, silently.
- engineer unattend, specialize Order 37 (free slot): run the script with
-StageOnly, copying the payload and the script to C:\PreInstall\engineering
while C:\Deploy exists (Path 119 chars)
- first logon Order 11: run it from C:\PreInstall\engineering, falling back
to C:\Deploy if staging did not happen
- the script (now in the repo, playbook/engineering/; it lived only on the
server): after the install, restore the Public desktop icon if it is
missing and the PC is not configured yet (no C:\TMP\DUU_Complete.log), and
register Active Setup so each later user gets the icon on their own
desktop when the Public one is gone. v3.1 puts the icon on the Public
desktop; clicking it (configure, from GE's V: share) deletes it for all
- playbook deploys the script to gea-engineer extra_engineering; the 650 MB
payload stays site content (documented in SITE-CUSTOMIZATION.md)
Tested in the win11 VM with the exact unattend strings and a stand-in
installer that skips the icon: specialize staged payload and script;
C:\Deploy deleted; first logon installed from the staged copy, restored the
icon and registered Active Setup. The per-user step as a new user: icon on
their desktop when the PC is unconfigured and the Public icon is gone; none
when configured or when the Public icon is there. Lint clean. Not deployed:
unattends go out only through scripts/deploy-unattend.sh.
Everything after OOBE needs the employee to log in, which happens off the
PXE network, so the floor waited forever at step 3 of 5. Now 3 steps:
WinPE, Windows setup installing apps, ready at OOBE (succeeded).
- both unattends: specialize Order 49 posts step 3 as succeeded
(Path 182 chars; lint clean). Only that line changed.
- Send-ImagingStage.ps1: StageTotal 3; steps above the total (the old
first-logon steps 4 and 5 still in the unattends) are logged and not
posted, so they cannot reopen a finished build
- winpe-status-push.ps1: total 3 for gea-standard / gea-engineer
- floor tile step 3: "Ready at OOBE"
Tested in the win11 VM against the test server: steps 2 and 3 post,
step 3 shows succeeded, step 4 is skipped. Not deployed: the unattends
go out only through scripts/deploy-unattend.sh.
Small Control JSONs have arrived empty or wrong more than once, and the old
check compared sizes only and skipped missing files. A copy of
Deploy/Control on the dev PC still holds two 0-byte JSONs.
Upload-Image.ps1 (tested on Windows PowerShell 5.1 in the win11 VM):
- refuses an unfinished MCL cache: exactly one <OS>-Production.tag (MCL
writes it last), no Control file newer than it, selected OS matches,
every Control/Tools JSON non-empty and valid
- checks payload against the SHA-256 MCL's own Control JSONs publish. MCL
resumes downloads onto old files (FileMode.Append) and then trusts any
file of the right size: the September junk-tail corruption. Offers to
delete the bad cache files so MCL downloads them again
- hashes every file into a manifest, robocopies, POSTs the manifest
- re-sends bad small files over HTTP (no proxy), large ones by deleting
the bad copy and robocopy /J (robocopy skips a same-size, same-time file
as "modified" even with /IS /IT)
Webapp:
- services/upload_verify.py hashes the upload on the server's own disk in
a background thread; reports missing, differing, bad-in-MCL-cache and
stray files; PUT /api/upload/file writes temp, checks hash, renames
- import of the upload folder only after a passed verification, with
every file unchanged (size, mtime, ctime, inode), no stray files, and
the target the OS allows (LTSC: Shop Floor only; else Standard or
Engineer; unknown OS: none). Import, verify and PUT share one lock
- paths matched case-insensitively against the disk, as Samba does
- import page shows the upload, MCL run, verification and stray files
with a remove button; record of each import kept on the image
Reviewed adversarially; all 11 confirmed findings fixed with tests.
42 new unit tests. VM test: clean upload; same-size same-mtime damage on
3 files repaired in one round; stray file blocks import until removed;
Shop Floor refused for a Pro upload; import records and clears the
folder; missing tag, 0-byte JSON, junk-head WIM and late Control file
each refused with nothing uploaded.
- Deploy shopfloor-setup with rsync instead of the copy module: the
~4 GB tree took about 4 minutes on every re-run because copy checksums
each file. rsync transfers only changes; changed only when it did.
- iPXE drops a trailing space, so 'Machine number:' ran into the typed
digits. The space now comes from a hex 20 setting.
- tests/vm: pxetest-xp is a permanent client (SeaBIOS, qemu32 32-bit CPU,
rtl8139, 1 GiB, IDE disk) for the i686 Clonezilla path; start.sh xp.
Only one client may run at a time across all three.
Boot menu: "Clonezilla: back up this PC" and "restore this PC" ask for the
machine number and boot Clonezilla (64-bit with Debian shim, or 32-bit
i686 by cpuid) with ocs_repository=smb3://clonezilla:<pw>@server/clonezilla/,
ocs_prerun fetching clonezilla-scripts/pxe-cz.sh and ocs_live_run running
it. The script validates the machine number, saves
<machine>-<YYYY-MM-DD-HHMM> (server time), writes pxe-meta.txt, runs
ocs-chkimg, and restores only that machine's images with -k1 -r after a
typed YES. The old entry is "Clonezilla: manual".
Playbook: clonezilla Samba user for the [clonezilla] share only, password
from clonezilla_smb_password (validated, filled into the menu
placeholder), scripts installed to the web root.
Webapp: Clonezilla backups page lists image folders and zips grouped by
machine number with serial/model/version/check, serial mismatch flag,
delete, move to another machine, streamed tar download, disk space bar
and an explicit keep-last-N retention with preview. Tests and a VM test
plan added.
Current Clonezilla is amd64-only. The menu now checks the CPU with iPXE's
cpuid --ext 29 (long mode): 64-bit CPUs get the current release (through
Debian's shim under Secure Boot), 32-bit-only CPUs (Pentium 4 without
EM64T, Pentium M, early Atom) get Clonezilla 3.2.0-5 i686, the last
32-bit release, from clonezilla-i686/. Without a cpuid command the 32-bit
build is used, which also runs on 64-bit CPUs; UEFI always gets the
64-bit build. A 'Clonezilla Live 32-bit' item forces it.
prepare-boot-tools.sh downloads it, the playbook deploys it.
Clonezilla Live runs a Debian-signed kernel whose modules (usb-storage,
uas, exfat) are signed by the same build key, so the menu now starts it
through Debian's Microsoft-signed shim with the iPXE shim command, the
same way Blancco uses Ubuntu's. prepare-boot-tools.sh extracts that shim
from the Clonezilla release (EFI/boot/bootx64.efi -> clonezilla/
shimx64.efi) and warns when the shim is not Microsoft-signed or the
kernel not Debian-signed.
Memtest86+ is unsigned and needed Secure Boot off; memory tests now use
the Dell built-in diagnostics (F12 at power-on, Diagnostics), noted in
the menu. The playbook removes the old memtest web dir. Nothing on the
menu needs Secure Boot off any more; the 'disable Secure Boot' screen
remains only as the fallback when a signed path fails.
H1 corp_admin_cidrs (UFW 22/9009 on the corp NIC from these only; refuses
when it would lock out the current SSH client) and corp_ssh_password_auth
(sshd drop-in, passwords only from the imaging subnet). Defaults keep the
current behaviour and print a loud warning.
H2 pxe-netplan-apply.sh: pre-flight refusal (pxe-MAC NIC holds the default
route, a DHCP lease or the route to an SSH client), run from the USB
before anything changes; post-apply checks (names, imaging IPv4, corp
IPv4 back within 30 s, SSH route) and a 120 s systemd-run auto-revert.
H3 restore() puts the previous NIC names back (two-phase rename via
temporary names, so swaps work) before generate + apply.
H4 a failed rename is fatal (restore + exit 1); the play drops the old
any-interface UFW rules only after verifying both NICs and their
per-interface SSH rules.
H5 first-boot checks the playbook exit code, retries at next boot (max 5),
writes /var/lib/pxe-firstboot.failed and /etc/issue + /etc/motd banners.
M1 build-usb.sh verifies iPXE files against the playbook pins, every wheel
against its own package's hashes, drops the unpinned pip fallback,
requires --site-vars or --no-site-vars.
M2 webapp restarts when the code, wheels or unit changed.
M3 ignore-carrier on the imaging NIC; ip_nonlocal_bind kept for the
apache/networkd boot race.
M4 menu suffix must match blancco_kver; missing kernel with no earlier
build is fatal; a live menu with the staged Blancco test entry is not
overwritten.
M5 build-proxmox-iso.sh is unsupported (hard exit); use build-usb --image.
M6 deploy-unattend.sh: password via SSHPASS env and sudo stdin only,
%q-quoted remote args, sha256 check of the live file; lint-unattend.py
reads PXE_SERVER/PXE_USER/PXE_PASS and uses sshpass -e.
M7 the playbook lints the unattend files before the force: yes copies
(lint-unattend.py ships on CIDATA as scripts/); pre-commit ACMR.
M8 Blancco initramfs build: no ignore_errors, pipefail, trap cleanup.
Also: Send-ImagingStage.ps1 deploy, carrier hook install tasks
(pxe-nic-carrier, networkd-dispatcher), L1 imaging-active hold
(allow_disruptive) and no iPXE downgrade below an Updates-page install,
L2 iso-volid, L3 kver pin read from the playbook, L4 cp -rL, L5 refuse
>= 4 GiB WinPE files, L6 PIP_DIR wipe guard, L7 per-package wheel hashes,
no online fallback, L8 check-updates.py (apt_pkg/cached compare, all
components, exit 2 on no debs / bad answers).
- pxe-nic-check: count 802.1X only from an authenticator (EAPOL type 0
with EAP Request/Success/Failure). iPXE sends EAPOL-Start on every NIC
open and Windows sends Start/Response, so a PXE-booting bay made the
verdict 'ge' and the watch stopped dnsmasq for the floor.
- pxe-nic-check: join the STP/PAE/LLDP/CDP/PVST+ multicast groups on
every capture socket, so CDP-only ports are heard on a plain NIC.
- pxe-nic-check --watch: if dnsmasq was (re)started inside the window,
re-check (~6 s) before stopping it; skip when already stopped.
New --state-cmd and --print-imaging.
- pxe-nic-carrier (+ pxe-nic-carrier.service): networkd-dispatcher hook
for the imaging NIC only. Carrier loss stops dnsmasq; carrier gain
restarts it through its guard if it was running or the hook stopped it.
- pxe-dhcp-hook.sh: clear only SMB (tcp 445/139) conntrack entries and
sockets, not all traffic for the client, which raced iPXE's first
HTTP fetch.
- GetPxeScript.aspx: every boot line has a failure branch (iPXE 2.0
accepts an unsigned EFI kernel and fails at boot under Secure Boot,
which aborted the script).
- Tests: client EAPOL, packet_mreq packing, watch race, shell helpers on
stub PATH, lab scenarios eapol-client, br-eapol-client, race, carrier.
- imaging: coerce client payload and stored session fields to str/int,
drop lists/dicts; tiles template guards pctype/current_stage; POST
rejects non-object bodies and non-string serials.
- nic_status: normalise every check-file entry; load nic.conf as-is with
a [nic] fallback; split interface= lists; keep the check's binding while
fresh, resolve wildcard sockets from the dnsmasq pin (never True for an
excluded NIC); mark vanished NICs down; include role bridges; stale or
future-dated files downgrade to warnings and drop evidence.
- Overview renders the network pill in its own try and shows
'Network: unknown' when status fails.
- ipxe_update: read every Authenticode signature (all WIN_CERTIFICATE
entries plus nested SPC_NESTED_SIGNATURE) with openssl; shim swap and
install need both Microsoft UEFI CA 2011 and 2023; install re-reads
each file (snponly ipxe.efi only); staged.json schema version, older
records shown stale and refused; uhci counts as native USB.
- backups describe their files; risky rollbacks need a confirm flag.
- mac_folders: guarded listdir, cached describe, configured TFTP root,
warning for a shim without ipxe.efi.
- evidence pack: years 2000..2100, 400 days inclusive.
- a11y role=group on quick ranges; progress bar stays full during 'shim'.
- tests/test_review_fixes.py: 29 unittest cases incl. both real shims.
- autoinstall: dead mirror + fallback offline-install, so the installed
base is exactly the ISO's; late-commands run offline-apt-install.sh on
CIDATA packages/ (local apt repo, one transaction, errors fail the
install, log /var/log/offline-apt-install.log); first boot no longer
runs dpkg -i; real archive sources written back at the end.
- download-packages.sh --iso: unpacks the ISO's default install source
into a docker image (pxe-iso-base:<version>) and downloads there: all
updates of the base, the explicit list (zip included) and the newest
pool packages with their closure; writes Packages + explicit.list and
checks the set offline in a fresh base container (--network none).
- offline-apt-install.sh: new; also used by that check (--simulate).
- build-usb.sh: ships Packages, explicit.list and the script; refuses a
set without them.
- tests/vm: pxetest-client-nosb (Secure Boot off, e1000e, SATA), one
client at a time.
Clonezilla 3.2.1-6 is gone from the mirrors (the download failed and the
unzip aborted the script before Memtest). Memtest moved its downloads to
download/v<ver>/, and 8.x names the x86_64 UEFI binary mt86p_<ver>_x86_64
with no .efi suffix; the script now finds it.
- webapp/requirements.txt pins all eight wheels with sha256, at the fixed
releases (lxml 6.1.3 for GHSA-vfmq-68hx-4jfw / PYSEC-2026-87, werkzeug
3.1.9 for GHSA-g6x2-hccm-hh4m).
- download-packages.sh fetches exactly that set for Python 3.12 / x86_64
into an emptied folder; PIP_DIR is honoured inside the container (own
mount), files are handed back to the calling user, USE_DOCKER=1 forces
the container on a noble host, an unknown package name now prints apt's
error instead of failing silently, and linux-modules-extra-7.0.* is gone
(it does not exist; 7.0 ships those modules in linux-modules).
- Playbook: the wheel installer refuses any wheel whose sha256 is not
pinned, needs exactly one wheel per pin, and removes an older installed
version by its RECORD first. The webapp copy is a full rsync every run
(was cp with creates: app.py, which never updated an install).
- scripts/check-updates.py: offline debs against noble security/updates
and the pins against PyPI + osv.dev; exit 1 when something needs
updating. Run it before building a USB.
Tested: installer in a Python 3.12 venv (clean install, upgrade over lxml
6.0.4 + werkzeug 3.1.8 leaves no old dist-info, re-run is a no-op, a stray
wheel is refused); check-updates.py flags the April bundle (94 security
updates) and passes the pins.
- Copy the whole webapp/ tree (minus __pycache__) in build-usb.sh and
build-proxmox-iso.sh, so config, services/, views/ and
blancco_evidence_monthly.py reach the server.
- sfdisk --append failure now aborts; the CIDATA partition is the one
starting exactly at the computed sector, and nothing that starts
before it is ever formatted (the old fallback hit the ISO's EFI
partition).
- Volume id and El Torito load sizes come from the ISO
(xorriso -report_el_torito as_mkisofs); grub.cfg must contain
autoinstall after patching.
- --image FILE SIZE|auto builds the same layout in a file without root
(mkfs.vfat -C + mtools, sfdisk on the file). Plus --payload-root,
--skip-enrollment/-bios/-drivers/-boot-tools, --yes.
- tests/vm: libvirt harness (pxetest-server, pxetest-client, isolated
pxetest-imaging network) with fetch-iso/build-image/create/start/
status/stop/destroy.
Every server has an imaging NIC (static, the only one that serves
DHCP/TFTP/HTTP boot/SMB) and a GE corporate NIC (DHCP client, SSH and the
webapp on 9009). The playbook guessed the PXE NIC and the installer put
172.16.9.1/24 on every en* NIC, so a two-NIC box could answer on the GE
network.
- New per-site vars pxe_nic_mac / corp_nic_mac (required, from
playbook/site-vars.yml or -e; template site-vars.example.yml), stable
names pxe_nic_name / corp_nic_name (default pxe0 / corp0, overridable),
pxe_server_ip, pxe_subnet, DHCP range. pxe_imaging_mac / pxe_ge_mac are
derived from them. The play stops before touching anything if a MAC is
unset, malformed, duplicated or not on the box (lists the box's NICs).
- Heuristic NIC detection and the single-NIC netplan task replaced by a
two-NIC netplan template (match macaddress + set-name), installed by
pxe-netplan-apply.sh: validate in a scratch root, back up, generate,
rename, apply, restore the previous config on any failure. cloud-init
network config disabled so it cannot rewrite it.
- dnsmasq: /etc/dnsmasq.d/pxe-nic-only.conf from the vars (interface=,
except-interface=, no-dhcp-interface=); dnsmasq.conf IP/range from vars.
- UFW per interface: imaging NIC 22, 67/udp, 69/udp, 80, 4433, 445, 139,
137/udp, 138/udp, 9009; GE NIC 22 and 9009 only. Old any-interface rules
and the stale 9011 rule removed after the new ones are in.
- sysctl drop-in: IPv4/IPv6 forwarding off; ip_nonlocal_bind so Apache
can bind the imaging IP before the address is up.
- Samba [global]: interfaces = 127.0.0.0/8 <imaging NIC>, bind interfaces
only. Apache 80 and 4433 on the imaging IP only; 9009 stays on all (TODO:
the webapp has no auth).
- autoinstall: DHCP only on en*, no static 172.16.9.1; apt geoip off and
offline-install fallback now that the corp NIC may have a lease.
- Blancco UEFI/Secure Boot: blancco_kver (7.0.0-34-generic) drives the
kernel, initramfs and modules builds instead of the running kernel;
outputs vmlinuz-ubuntu-70, kexec-initrd-70.img, kmod-<kver>.tar.gz as the
menu and blancco-init.sh expect; creates: guards keyed on the full kver.
Ubuntu's shimx64.efi.signed published as blancco/shimx64.efi.
download-packages.sh adds shim-signed, linux-generic-hwe-24.04 and the
pinned linux-image/modules/modules-extra for BLANCCO_KVER.
Rendered with the live MACs and names, netplan generate output is
identical to the live config; pxe-nic-only.conf and the Samba binding
match live directive for directive.
A PC whose imaging failed and is re-imaged gets the same IP, and the
server still held state from the dead run, so net use failed (System
error 53). The April fix (18537ac) did not cover it:
- The dnsmasq hook only acted on 'add'/'del'. A re-image comes back to the
existing lease, which dnsmasq reports as 'old', and only when the
hostname/MAC changed. Now dnsmasq sets script-on-renewal, and the hook
clears state on 'old' when the request comes from PXE firmware or iPXE
(vendor class PXEClient), i.e. the machine just network-booted. A
Windows lease renewal (MSFT 5.0) is left alone.
- The 1 h conntrack timeout never applied after a reboot: systemd-sysctl
runs before ufw loads nf_conntrack, so the keys did not exist and the
5-day default stayed. /etc/modules-load.d/pxe-conntrack.conf loads it
first.
- smb-soft-reset.sh flushed neighbours on br-pxe, which no longer exists.
Live 2026-10-14. Hook tested on the server with an unused IP: network-boot
renewal clears, Windows renewal and no vendor class do not, add clears.
pxe-nic-check classifies each NIC from what it hears: LLDP/CDP/802.1X
from managed switches, foreign DHCP offers and their domain (one
DHCPDISCOVER on the imaging NIC; the GE NIC's networkd lease is read
instead), STP and ARP hosts for information. A kernel BPF filter keeps
bulk SMB traffic out of the capture loop; bridge ports are listened on
too, since LLDP and EAPOL stop at the port.
- dnsmasq drop-in runs --guard as ExecStartPre: exit 1 on a ge or
foreign-dhcp imaging verdict, fail open (loudly) on bugs or no link
- pxe-nic-check.timer runs --watch every 5 min: stops dnsmasq on a bad
imaging verdict, never starts it
- /run/pxe-nic-status.json (0644, atomic) for the webapp
- playbook installs it all and writes /etc/pxe-server/nic.conf from
pxe_imaging_mac, pxe_ge_mac, pxe_corp_domain_suffixes,
pxe_nic_allowed_switches
- unit tests plus a no-root network-namespace lab (veth pairs, a raw
socket DHCP/LLDP peer and real dnsmasq); docs/pxe-nic-guard.md has the
operator notes and the two floor tests
a94c203 added a RunSynchronous Path of 417 chars to both std/eng answer
files. Windows caps it at 259 and then rejects the WHOLE answer file, so
an Engineer build stopped at 'Windows could not parse or process unattend
answer file'. The live files were restored to their pre-a94c203 versions
the same day.
- The status steps now fit: Send-ImagingStage.ps1 -Persist copies itself
out of C:\Deploy (no cmd copy chain), short PowerShell switches, default
arguments left out. Longest Path is 168 chars; lint-unattend passes.
NOT deployed: these files must go live only through deploy-unattend.sh.
- scripts/deploy-unattend.sh: the only route for an answer file to a PXE
server. Lints first and refuses on any error, backs up, swaps
atomically, lints the live copy and restores the backup if that fails.
- .githooks/pre-commit (git config core.hooksPath .githooks): refuses a
commit whose staged unattend fails the linter. Verified by planting a
352-char Path (blocked) and against the exact a94c203 file (rejected).
- Boot menu: also accept SecureBoot read as hex 01, and show the raw values
in the warning so a wrong read can be diagnosed from the screen.
- winpe-status-push.ps1: check-bios.cmd stages the firmware update before
the imaging menu, but startnet's 'BIOS firmware update' push only comes
at the end of WinPE, so builds showed a plain 'Applying image' for the
whole apply. The first push now reports the staged update when
X:\bios-fired.flag exists, and the floor shows 'Applying image - BIOS
update staged' (do not power off during the next reboot). Also: CIM
fallback for the serial where wmic is missing, and standard/engineer
always report 5 steps.
- Network pill: cdc_ncm (e.g. the WisdPI RTL8157 5G adapter) keeps link
speed in a 32-bit bits/s field, so 5 Gb/s reads 705 Mb/s; show the real
speed with a note.
Tested in the win11 VM (with and without the BIOS flag, and the late
startnet push).
A 'Network' pill next to 'Services' lists each physical NIC with its role
(imaging / GE network, from /etc/pxe-server/nic.conf or dnsmasq's
interface= pin), link, speed, addresses, and whether dnsmasq serves
DHCP/PXE on it. It reads pxe-nic-check's verdict file
(/run/pxe-nic-status.json) when present, and shows a red banner when the
imaging NIC is on the GE network or a network with another DHCP server,
or when dnsmasq answers DHCP on a non-imaging NIC. Without the file it
shows link and addresses only. Read-only.
Live since 2026-10-14 for every client (all UEFI clients now run iPXE 2.0):
- WinPE skips the 'Secure Boot must be ENABLED' warning and its 5 s wait
when iPXE reads Secure Boot as on (${efi/SecureBoot:uint8}); otherwise,
or on an iPXE that cannot read it, it warns as before.
- Blancco and Blancco debug start through Ubuntu's shim (iPXE shim
command), so they run with Secure Boot on; an iPXE without the command
boots the kernel directly as before. The separate Secure Boot TEST entry
is gone.
- Blancco has its own menu section; only Clonezilla and Memtest stay under
'Secure Boot must be DISABLED'.
The playbook now installs playbook/ipxe/GetPxeScript.aspx instead of an
inline copy that had drifted from the live menu. Checked that the Blancco
Releases page still parses it (one item blancco, :blancco_efi block).
Live since 2026-10-14: dnsmasq hands every UEFI client ipxe-shim.efi
(iPXE's shim, dual-signed under Microsoft UEFI CA 2011 + 2023), which
loads ipxe.efi from the TFTP root, now the iPXE v2.0.0 snponly build
(firmware network and USB drivers; the full build kills USB keyboards).
The per-MAC test tags and folders are gone; the old Broadcom-signed
1.21.1 build stays as ipxe-bcm.efi, the rollback.
- playbook: pinned sha256 vars, the new dnsmasq default line, and a
verified, non-optional task that installs ipxe-shim.efi, ipxe.efi
(snponly), ipxe.pxe and ipxe-bcm.efi (was: copy ipxe.efi and the never
served undionly.kpxe, errors ignored).
- scripts/fetch-ipxe.sh downloads the pinned tarball and shim into
boot-files/ for the USB build; README/SETUP list the step.
- iPXE page: ipxe.efi is staged from snponly.efi, a build that drives USB
itself is refused, and the snponly-shim/snponly root files are gone.
Tested: the playbook task against the real release files (installs the
live hashes; a tampered file stops it before anything is replaced),
ansible --syntax-check, staging, and the full-build refusal.
The floor's 8 stages are the shopfloor pipeline, and only shopfloor-setup
reports after WinPE, so every gea-standard / gea-engineer build sat at
1/8 forever with no type.
- winpe-status-push.ps1 reads the image from the Z: share mapping when no
PCTYPE is set and reports those builds as 5 steps.
- Send-ImagingStage.ps1 (deployed to _shared\Applications\extra\pxe-status)
posts steps 2-5 from the unattend: start and end of specialize, first
logon, and setup complete (succeeded). It is copied to
C:\ProgramData\PxeStatus first, because the bulk-enrollment task deletes
C:\Deploy before first logon. Best-effort: always exits 0.
- Floor rows use Standard/Engineer labels and show the image as the kind.
Tested as SYSTEM in the win11 VM (PowerShell 5.1): the exact unattend
command line, stages 3-5 against a test server, a dead port (exit 0), and
the image regex.
Staging fetches ipxe-shimx64.efi from the ipxe/shim release (sha256 from
GitHub), and uses it instead of the tarball's 2011-only shim when it is
signed under the Microsoft UEFI CA 2023 and trusts the same iPXE CA;
otherwise the tarball shim stays, with the reason shown. snponly.efi and
snponly-shim.efi are managed alongside ipxe.efi and ipxe-shim.efi, each
iPXE build says whether it drives USB itself (kills USB keyboards), and
tftp-unique-root per-MAC folders are listed read-only.
Evidence pack gets a Last 3 months quick range (three full calendar
months). Opened menus are nudged back inside the window, so header
popovers no longer start off-screen on phones or a wrapped header.
Buttons and filter pills inside a menu keep their own styling.
Evidence pack takes ?from=&to= (inclusive, Eastern time) with From/To
pickers and quick ranges; ZIP is named ENT-056_<site>_<from>_to_<to>.
Month keys show as 'September 2026'. Hardware gap rows on the to-do
use the shared one-line grid with chips for the missing pack or BIOS.
From the design review:
- Floor rows reserve 150px for actions instead of 190px; stage text is
one line with the full text in a tip (safety text still wraps); the PC
type hides first; rows stack only below 960px of panel width, with the
actions on the status line instead of their own row.
- Tips also show on keyboard focus.
- Placeholders use the muted token (dark-theme contrast was 3.6:1).
- Save bar focus ring uses the save bar's accent (2.8:1 before).
- Unattend: a command's '...' menu stayed open after Move up/down and
covered the next row's menu, so the next click could stage removing the
wrong command. The menu now closes and focus returns to its button.
- '...' menus close on an outside click, on Escape and when another
opens (static/app.js), for every page.
- Drivers: Remove asks to confirm again, as before the redesign. Long zip
paths in a row body wrap (the page scrolled sideways at 390 px).
- Rollback review returns to the page it was opened from (Dell page or
Updates overview), for both Cancel and the rollback itself.
- To-do: the ServiceNow period offers a year again (366 days).
- Imaging floor: a search that matches a log-only bay opens that group;
its open state survives the live refresh. 'Clear all' states the
number of session files it deletes, hidden finished ones included.
- Dashboard: the image key pattern is valid under the browser's v flag
(client-side validation was silently skipped).
- Import: allowed_import_source() compares real paths, so
'<upload dir>/../..' is refused. This path traversal predates the
redesign; the import moves the source tree, so it mattered.
Blocker
- /reports/export returned 500: REPORT_EXPORT_RANGES and _months_ago were
deleted with the old reports route. Restored.
Major
- A session without stage_index/stage_total took down /, /imaging and
/imaging/tiles (int filter on Undefined). The templates default both;
saving a LAPS password for a session that no longer exists is refused
(it created a bare session).
- /reports/todo parsed every report four times (23 s at 700 reports).
blancco_report.parse() now caches by path, mtime and size: 0.09 s.
- Stored XSS through the session serial in the Clear confirm (the
browser decodes ' before running an inline handler). The message
moves to data-confirm, and update_session() only accepts serials of
letters, digits, '.', '_' and '-' (400 from /imaging/status otherwise).
- Reports without a document id and with odd characters in the file name
got a key the mark routes rejected, so they could never be cleared.
report_key() now always produces a valid key.
Minor
- failure_note() handles a list-valued erasure target (500 on the report
and to-do pages).
- _recently_marked() tolerates naive, 'Z' and non-dict state entries.
- Key regexes use fullmatch (a trailing newline was accepted).
- Unreadable reports are no longer counted as to tick / to destroy; the
reports page already flags them as Unreadable.
- Rollback review recounts restorable files after reading the backups.
- Enrollment: the startnet fallback package (no or empty ppkg.conf) and a
quoted SOURCE_PPKG are protected too.
- Backup upload overwrite check is case-insensitive (SMB share) and
replaces the existing file rather than adding a case twin.
- Overview polling keeps the last good rows on an error and pauses while
the tab is hidden. A shown or typed LAPS password survives the floor's
refresh.
- hardware_dismiss only redirects to same-site paths; tiles ?limit is
clamped; the Updates overview survives an unreadable Dell scan or iPXE
state; describe() uses aversion(); a client cannot label its session
'inferred'.
Restyled to the same row pattern (one line, details in an expandable
body, one action plus a menu, tokens for light and dark):
- hardware coverage, audit log (family chips, destructive actions in
red, client-side filters, first 300 rows), Clonezilla backups,
enrollment packages (active package marked and undeletable, MCL flags
with a visible legend)
- image Overview, OS and packages, Unattend (sticky save bar), Import
- Dell, iPXE, Blancco and BPRT update pages; Dell history links to the
rollback review instead of rolling back on one click
- boot menu editor (tabs, rows with a menu, lint chips that jump to the
line, tokens instead of hard-coded dark colours) and the imaging
session detail page
Fixes found on the way:
- BPRT batch Activate never asked to confirm: a hidden input named
'confirm' shadowed window.confirm in the inline handler, so the form
submitted without a prompt. It now calls window.confirm.
- Backup upload refused when a backup with that name exists, unless
'Replace the existing backup' is ticked (it overwrote silently).
- app.js: em dashes in comments replaced (plain ASCII).
Reports (/reports):
- Grouped by month of the erasure date inside each report (not the file
date), newest first; older months fold. Each month header shows the
evidence pack state (builds on the 1st, to upload with Download, or
uploaded).
- One line per report: erasure date, system serial, model, drives,
result ('Successful, fallback' when a failed attempt fell back to
another standard), cloud and ServiceNow state. The row body lists every
drive and erasure attempt with the failure reason.
- Search by system serial, model, drive serial or file; filters for
failed, not in cloud and to tick. 'To tick' uses the to-do's 30-day
window; older unticked reports read 'Not ticked'.
- Export and the evidence pack download moved into header popovers.
Report (/reports/view/<file>):
- Header with serial, model and the per-drive result; a notice names
the drive to destroy until its COD is recorded; Copy ServiceNow note.
- One card per drive with every attempt, a step track and the failure
reason; steps and error counts fold. System, customer and provenance
(integrity signature folded, copyable) on the side. Sizes in decimal
units, as on the list.
- blancco_evidence.failure_note() is shared with the ServiceNow list.
To-do (/reports/todo):
- The ServiceNow tick list is now its first section (period and show
ticked filters); /reports/servicenow redirects there.
- Five counters; one line per item with details in the row body.
- COD recorded, Mark uploaded and Resolved confirm first (the attribute
quoting is safe for JSON) and can be undone for 24 hours from the
page. The header total includes hardware gaps, matching the sidebar.
- GET /updates/: one card per updater (Dell, iPXE, Blancco, BPRT) with
its state, a Check now / Open action and details in accordions; the
shared download cap is a small popover.
- Dell applies list: one line per apply titled by its first change
('Latitude 5350 pack A11 to A12', model looked up in the maps) and
'and n more'; the chip says how many written files are unchanged since
the apply. The body lists every change, skipped item, error and file.
- dell_updates.file_status(): per file an apply wrote, unchanged /
changed / missing by content hash, the same test rollback uses.
describe(), history_view() and rollback_preview() build on it.
- GET /updates/dell/rollback/<id>: review a rollback before doing it,
with the files that would be restored and a diff of each. The POST
route is unchanged; both forms can return to the overview.
- One line per entry: position, model tokens, pack, OS, size and status
chips; family, friendly name, osId, folder, zip state, Dell hint and
the verbatim lint findings are in the row body.
- Status chips compare the staged map with the saved one: New, Edited,
Moved from #n, Removing, the worst current finding, and 'Clears on
save' / 'Clears n' when the staged changes resolve this entry's
findings. The toolbar and a sticky save bar show lint now and after.
- Removing an entry is staged with Undo remove instead of deleting the
row; Undo edit puts an edited entry back.
- Reordering moved into the row menu (Move up / down / to position).
The Move dialog shows the order after the move and re-lints the staged
map with the move applied, listing any new errors before staging it.
- An edit that changes the tokens but keeps the old friendly name is
flagged in the row and in the review step.
- Filters: Win11/Win10 with counts, 'Findings or unsaved', 'Zip
problems' (missing, trap and case); a status key explains the chips.
- _image_header.html: new header and underline tabs for every image tab;
the shared review modal and diff use the design tokens.
Imaging floor (/imaging, /imaging/tiles):
- One line per PC: status chip, serial with machine number, hostname
and IP, stage text with a segmented stage track, one action plus a
'...' menu. Everything else (description, error, timestamps, Intune
ID + QR, Set category / ARTS links, LAPS QR, log tail) is in the row's
expandable body.
- Order: failed, needs action (stage 7), in progress, done; log-only
bays fold into one collapsed group. imaging_status.floor_state() and
floor_order() compute this.
- Status filter buttons with live counts next to the text filter. SSE
refresh, expanded-row memory and LAPS handling are unchanged.
- Stage text is never truncated; BIOS updates and erasing are bold.
Overview (/):
- Services popover, technician to-do line, the six most urgent floor
rows (?limit=6 partial, refreshed every 15 s), and image cards with
lint warnings, map errors and missing zips as separate chips. New,
clone, delete and unattend stay on the cards' menus.
Copy buttons are now handled once in static/app.js, with a fallback for
plain-http pages where navigator.clipboard is unavailable.
- static/pxe-ui.css: the redesign's colour tokens (light and dark),
shell, chips, stage track, expandable rows, menus and buttons. The
legacy tokens (--bg-card, --border, --text-light, ...) alias onto the
new ones so pages not yet redesigned blend in.
- base.html: new sidebar (Operate, Images, Updates, Tools) with the GE
Aerospace logo on the brand blue, to-do count, theme toggle with
aria-pressed, and a menu button on narrow screens. The to-do banner is
now a one-line strip. Skip link and <main> landmark added.
- _ui.html: Jinja macros for chip, track, the actions menu and the page
header.
- Report view: the hero summarised only the first erasure, so a drive whose
Purge failed and Clear fallback succeeded showed Failed while the list said
Successful. It now uses the per-drive outcome (report_result) and lists the
standards that wiped each drive.
- To-do and ServiceNow: reports with no usable document id could never be
ticked or cleared (400). They are now tracked by file name (file:<name>).
- Imaging: the LAPS password field is masked, with a Show/Hide toggle.
- Imaging: succeeded sessions not updated for ACTIVE_WINDOW_HOURS are hidden
from the dashboard (failed and in-progress always stay), with Show all.
- Enrollment: the active SFLD package can no longer be deleted, which left
ppkg.conf pointing at a missing file; its row is marked Active.
- Import: confirm before importing, saying upload-folder sources are moved
and same-name files in the image are overwritten.
