mirror of
https://github.com/mulgadc/spinifex.git
synced 2026-11-01 16:59:45 +00:00
Josh Siviour
488f1c7bf9
* feat: build a separate SQS store microVM image Package the go.mod-pinned nats-server, ext4 support and a clean ACPI power-down into a microVM kernel and initramfs of their own. The build checks the packed archive's contents and refuses a nats-server whose version differs from the pin. The image does not start the store and nothing installs or releases it; the LB image is unchanged. * fix: publish the SQS store image only after its gates pass Stage the kernel and initramfs and move them into place only once the content, size and guest-RAM gates pass, so a failed build cannot leave a pair make takes as up to date. The old initramfs is removed first and the new one lands last, so an interrupted publish never leaves a mixed pair. Check the module list without piping into grep -q, which under pipefail reported a present module as missing. * feat: start the SQS store guest on its checked data disk The SQS store image now prepares, runs and stops a dedicated JetStream server. It reads a narrow fw_cfg contract (data disk serial, listen address, one user NKey public key), refuses anything absent, malformed or ambiguous, has the pinned server test the config, then checks and mounts exactly one matching ext4 disk. It never formats and has no ephemeral fallback. The server listens only on the supplied guest NIC address, admits only that NKey and syncs every write. Shutdown stops the recorded run process first, whether it is the server or a start not yet exec'd, and runs the generic read-only remount only once it is gone. A store that survives SIGKILL leaves every mount alone. * fix: report the SQS store as prepared, not listening, before it starts * feat: launch the SQS store guest on its own versioned subject An explicit SQS store kind carries an existing system-account data volume and the one NATS user public key the store accepts. It is served only on system.LaunchSQSStore.v1, so a daemon without this code has no responder and the launch fails instead of booting a load balancer without its disk; the legacy launch path refuses the kind on both send and receive. Every refusal happens before capacity, an ENI or a VM record exists. The volume goes through the normal EBS mount and terminate detach and is never deleted on termination. The store boots its own image with CPU mitigations on, gets only its network, disk serial, listen address and key, and cannot be relaunched after a host restart yet. * feat: ship the SQS store microVM image in releases and installs Release builds now build the SQS store image and stage it beside the load balancer's as microvm-sqs-store/, so an older installer that copies microvm/* without -r still works. setup.sh installs it to /usr/share/spinifex/microvm/sqs-store/, where the daemon looks for it. A release without a complete image never leaves an older one usable: staging drops a previously staged image, and the installer removes an earlier installed pair before warning (missing) or failing (incomplete). Only the two image files are removed. The initramfs is installed last, so an interrupted install leaves no bootable mismatched pair. * feat: restart the SQS store guest on its launching node Persist the non-secret guest config of an SQS store instance (launching node, data volume, listen address, client public NKey and rendered network config) in the instance record, and rebuild its fw_cfg files from it on relaunch. The relaunch refuses on any other node, or when the record is missing or malformed, before writing any file; the data volume is remounted through the normal EBS path and kept on terminate. Cross-node recovery now never selects or claims an SQS store record, so no other node spends retry budget on it or relaunches it. Same-node restore is unchanged, as are load balancer restarts. Older daemons know neither the exclusion nor the record field, so this does not prevent a cross-node relaunch by a mixed-version cluster. * feat: hold prepared system volumes in creating until released Add an internal CreateVolumeForPreparation path that writes a new volume's first record as creating, and CompleteVolumePreparation to release an unattached creating volume in the caller's account to available. Ordinary CreateVolume is unchanged. Attach and delete already refuse a creating volume. ModifyVolume, ForceDetachVolume and CreateSnapshot now refuse it with IncorrectState so nothing can expand, release or capture a volume mid-preparation. * feat: prepare fresh SQS store volumes with an ext4 filesystem The SQS store guest never formats a disk, so a new volume gets its filesystem on the host. The preparer checks for mkfs.ext4 and qemu-img, builds an ext4 image in private scratch, creates the volume in "creating", publishes it on this node, copies the image with the NBD import helper, unpublishes, and only then releases it as available. It never writes a volume it did not just create. A volume is deleted only when no export can still be live: after a confirmed unpublish, or when the export was not NBD. Any publish error, a failed or ambiguous unpublish, or a failed release retains the volume and reports its ID. A publish refused as in-use or already-exists sends no unpublish, since another export may exist. Cleanup ignores caller cancellation but is bounded at 30s. The raw-image copy moves to runtime/compute/nbd unchanged, and the domain receives it as an injected writer. The volume service gains a discard path for an unattached volume still in "creating". * fix: report unknown state and match qemu-img lookup in volume prep A failed release may still have made the volume available, so the retained-volume error now says its state is unknown rather than claiming it is still creating. The copy runs qemu-img by name through PATH, so the pre-allocation check now uses a PATH lookup for it too. mkfs.ext4 keeps the sbin fallback because it runs from the resolved path. * docs: describe the unwired SQS store system instance Records what the guest image, versioned launch, same-node restart and fresh-volume preparation provide, that nothing calls them, and what is untested or missing before production use.
Spinifex Scripts
Node installation, dev-environment lifecycle, guest image builds, and verification tooling for Spinifex.
Install and node setup
| Script | What it does |
|---|---|
setup.sh |
The production binary installer, served from install.mulgadc.com. Installs deps, users, sudoers, the host firewall policy, files, systemd units, logrotate and udev rules, and loads openvswitch, geneve, vhost_net and tun, listing them in /etc/modules-load.d/spinifex.conf for later boots. It masks firewalld wherever it is installed, since the Spinifex nftables policy is the only supported node firewall. An unset --firewall keeps the mode recorded in /etc/spinifex/firewall/mode, and otherwise arms the policy when firewalld was active. Where systemd-resolved is installed but not running, resolvconf is absent and NetworkManager runs, it hands host DNS to resolved: it saves /etc/resolv.conf once to /var/lib/spinifex-reset/resolv.conf.orig, links /etc/resolv.conf to the resolved stub, and stops unless resolved has upstream servers and the package repository host still resolves. It links /usr/bin/spx to /usr/local/bin/spx where sudo's PATH leaves out /usr/local/bin, and takes the nbdkit plugin directory from nbdkit --dump-config, stopping when nbdkit cannot say. SETUP_STAGES runs a subset; ISO_BUILD=1 runs it inside the ISO builder's chroot. |
setup-ovn.sh |
Bootstraps a compute node for OVN VPC networking: installs OVN/OVS, creates br-int, configures the WAN bridge (bridged, DHCP, or routed-NAT uplink), sets the chassis identity, and applies overlay sysctls. It runs OVS and OVN as root, setting OVS_USER_ID and OVN_USER_ID where /etc/sysconfig holds them, and hooks the socket-permission helper onto ovs-vswitchd, ovn-controller, openvswitch-ipsec and, on EL, ovn-northd. --management also starts the OVN central services, writes their ovn-ctl flags to /etc/default/ovn-central (Debian) or /etc/sysconfig/ovn-northd (EL), and records the role in /etc/spinifex/ovn-role. With clustered databases it returns only once this node's NB and SB members are back in their RAFT clusters under a known leader, and fails after OVN_RAFT_REJOIN_WAIT seconds (default 60) otherwise, so reruns one node at a time never take two members of a cluster down. A later run without --management refuses while that file, or the OVN_CTL_OPTS line older nodes carry in /etc/default/ovn-central, records the role. It adds maxsize 100M to the OVS and OVN logrotate files, plus rotate 5 where the file sets no count, and rotates as root, matching the daemons. On a host whose IPsec stack is libreswan it enables, stops and masks openvswitch-ipsec.service and ipsec.service for the daemon to unmask when IPsec is on, unless openvswitch-ipsec.service is already running; elsewhere it enables and starts openvswitch-ipsec.service, unless the daemon has masked it to keep IPsec off. The links it creates persist through systemd-networkd where that is enabled, and otherwise through /etc/spinifex/hostnet.json, which spinifex-hostnet.service (spx admin hostnet apply) replays at boot and on every OVS start, starting ovn-controller again when an OVS stop took it down; the first run records the choice in /etc/spinifex/hostnet-backend. Where NetworkManager runs, it leaves the links Spinifex owns unmanaged. |
install-node.sh |
Forms a multi-node cluster across hosts that are already running Spinifex: rebuilds the OVN database in clustered form, runs spx admin init on the first host and spx admin join on the rest, then verifies. Before touching any host it refuses hosts of different OS families (distro_os_family) or spx versions, and --ipsec on when any host uses libreswan. The scripted form of install-multi-node. Destroys the joining nodes' logical network state and crypto — see its --help. |
dev-install.sh |
Full local dev setup through the production installer: builds from source, packs a tarball, runs setup.sh, initialises the cluster, and starts services under systemd. |
build-oracle-appliance.sh |
Turns an Oracle Linux 10.2 system reachable over SSH with sudo into a sealed gold image, then powers it off. It updates OL and reboots into the newest UEK R8, sets SELinux permissive, installs Spinifex without initialising it, and leaves spinifex.target and the NATS watchdog disabled. It creates the spinifex login with a default password that must be changed at first login, installs appliance/firstboot.sh as spinifex-appliance-firstboot.service, and wipes the system's identity (machine-id, SSH host keys, OVS chassis, NetworkManager profiles, logs). On each unit's first boot that service puts the first physical NIC with carrier on a DHCP br-wan, names the host spinifex-<mac>, runs setup-ovn.sh --management and spx admin init --nodes 1, and shows the IP and login on the console and in /etc/motd. --channel, --version and --tarball pick what to install. |
clone-deps.sh |
Clones or updates the viperblock and predastore checkouts alongside spinifex for cross-repo development. |
lib/distro.sh: host-OS capability probes
The node scripts select host-OS behaviour by probing for the capability, not by reading /etc/os-release; only setup.sh's supported-OS gate keys on the OS ID.
Each probe prints its answer, or names every candidate it tried on stderr and returns 1, which it also does when it finds both candidates of a pair.
| Probe | Answers |
|---|---|
distro_pkg_manager |
apt or dnf |
distro_ovs_unit |
openvswitch-switch.service (Debian, Ubuntu) or openvswitch.service (EL) |
distro_ovn_central_layout |
split (Debian's ovn-central plus separate NB and SB units) or lone-northd (EL's single ovn-northd.service) |
distro_ca_trust |
the trust-store update command and its anchor directory |
distro_ovs_logrotate |
/etc/logrotate.d/openvswitch-switch or /etc/logrotate.d/openvswitch |
distro_ovn_logrotate |
/etc/logrotate.d/ovn-common or /etc/logrotate.d/ovn |
distro_os_family |
the OS family of the /etc/os-release text on stdin: debian for Debian and Ubuntu, otherwise the ID |
distro_has_libreswan is a predicate rather than a probe: it succeeds when libreswan's IKE daemon, /usr/libexec/ipsec/pluto, is installed.
EL's openvswitch-ipsec drives libreswan even beside strongSwan, so finding both is not an error.
distro_missing_ovs_deps <unit-file>... is a check rather than a probe: it prints each OVS or OVN unit the given unit files depend on or order against that is not installed, and fails only when a file cannot be read.
setup-ovn.sh runs it against the installed Spinifex units and stops when one is missing, because systemd silently drops ordering against a unit that does not exist; a missing openvswitch-ipsec.service is only a warning.
setup-ovn.sh, uninstall-spx.sh and install-node.sh source it from lib/ beside themselves, then from /usr/local/share/spinifex/lib/distro.sh, where setup.sh installs it, and stop naming both paths when neither exists.
setup.sh is piped from the install endpoint before the tarball is downloaded, and install-node.sh --wipe copies node-reset.sh alone to each host, so those two carry inline copies of the probes they need between # --- BEGIN distro.sh probes --- markers.
lib/distro_test.sh fails when an inline copy differs from the library.
Go code uses spinifex/runtime/host/hostos for the same purpose.
System packages
setup.sh installs APT_RUNTIME_PACKAGES with apt, or on Oracle Linux OL_RUNTIME_PACKAGES with dnf, choosing by distro_pkg_manager.
Before the dnf install it enables ol10_UEKR8 and ol10_kvm_utils, excludes qemu-kvm* and qemu-img* from ol10_appstream so QEMU always comes from KVM Utilities, and writes /etc/yum.repos.d/spinifex-nfv-sig.repo for OVS 3.5 and OVN 25.03 from the CentOS NFV SIG.
That repo carries an includepkgs= allowlist, and its key is refused unless its fingerprint matches NFV_SIG_KEY_FPR.
It then installs kernel-uek-modules for the running kernel, stopping with a reboot instruction if that version has left the repositories, and loads openvswitch, geneve, vhost_net and tun.
make install-system takes the same dnf path, adding the build packages from CodeReady Builder.
| Script | What it does |
|---|---|
lint-apt-packages.sh <image> |
Resolves the apt list against a Debian or Ubuntu image's real indices with apt-get install -s. |
lint-dnf-packages.sh [image] |
Runs setup.sh's Oracle Linux repository setup in an oraclelinux:10 container, resolves the dnf list, and fails unless QEMU comes from KVM Utilities and OVS and OVN from the NFV SIG. |
check-glibc-ceiling.sh <ceiling> <elf>... |
Fails when an ELF file needs a glibc symbol version newer than the ceiling, naming the symbols. build/Dockerfile.distro runs it on spx and the nbdkit plugin with 2.34, so the one release tarball keeps loading on every supported OS. |
Teardown
Both ship to /usr/local/share/spinifex/ on every node, so they are available on a host installed with curl … | bash and not only from a checkout. They differ in what they take away:
| Script | What it does |
|---|---|
node-reset.sh |
Clears state and leaves the node installed: stops services, destroys guests, wipes /etc/spinifex, /var/lib/spinifex, /var/log/spinifex and the OVN databases, and removes the OVS bridges. On a libreswan host it deletes the ovs_cert_* and ovs_certkey_* entries ovs-monitor-ipsec imported into libreswan's NSS database, leaving other identities and /etc/ipsec.d/*.conf alone. It stops OVS and OVN central by the unit names the host's packages use. It leaves the NetworkManager unmanaged-device config, which the next setup-ovn.sh run rewrites, so NetworkManager cannot take a NIC back in between. The node is then ready for a fresh spx admin init. --keep-data preserves /var/lib/spinifex; SPX_RESET_CONFIRM=reset runs it unattended. |
uninstall-spx.sh |
Removes the software — the inverse of setup.sh. Delegates the state teardown above to node-reset.sh rather than reimplementing it, then removes the binary, the nbdkit plugin, the systemd units and slices, the service users, the sudoers grant, the sysctl/logrotate/chrony/udev drop-ins, the systemd-networkd units, the host firewall policy and its loaded nft table, the Spinifex systemd-resolved drop-in, and the Spinifex CA from the host trust store. It unmasks firewalld and leaves it disabled. Where setup.sh handed DNS to systemd-resolved, it restores the saved /etc/resolv.conf and turns resolved off. Verifies afterwards and exits non-zero naming anything that survived, then lists the host settings setup.sh changed that it leaves in place, such as the dnf repositories, OVS_USER_ID, masked IPsec units and the SELinux mode. |
node-reset.sh finds guests by process name, qemu-system-* or qemu-kvm, owned by the daemon's service user, and waits for them to exit before it deletes any state.
node-reset_test.sh checks that match against real processes, and that reset-dev-env.sh uses the same pattern.
uninstall-spx.sh keeps /var/lib/spinifex and every package by default — uninstalling software should not destroy data, and a curl … | bash install lands on hosts that were already using openvswitch, libvirt and chrony. --purge-data and --purge-deps opt into each, and both confirm separately even under --yes (SPX_PURGE_CONFIRM=destroy for unattended data removal). --dry-run prints the whole plan and touches nothing.
--purge-deps removes with apt and is refused on dnf hosts, where its fixed list cannot tell what setup.sh installed from what the host already had, and dnf would also remove every package that requires them.
It refuses on a multi-node cluster, and --yes does not override that. Uninstalling one node of a running cluster strands its guests and leaves the survivors reconciling against a chassis that will never answer. Run spx admin cluster shutdown first, then the uninstaller on each node.
Dev environment
| Script | What it does |
|---|---|
reset-dev-env.sh |
Single-node reset: stops services, wipes /etc/, /var/lib/, /var/log/ and /run/spinifex, rebuilds, reinstalls via dev-install.sh, and relaunches a smoke-test instance. Preserves the existing node/network topology from spinifex.toml. |
dev-env/*.conf |
Topology profiles for reset-dev-env.sh --profile=NAME, declaring which bridge carries the wan, lan and vpc planes: single-nic (all collapsed onto wan), two-nic (vpc folds onto lan), three-plane (a NIC each). |
Verification and benchmarks
| Script | What it does |
|---|---|
smoke-test.sh |
End-to-end sanity check on a running node: imports an SSH keypair and the Ubuntu AMI, then launches an instance. --create-vpc builds the VPC, gateway, subnet and routes first and tears them down after; --nodes N launches through a spread placement group so each instance lands on a different physical server. TEST_GPU=1 adds GPU passthrough checks. It runs the AWS CLI by absolute path, since sudo's secure_path leaves out /usr/local/bin on Oracle Linux. |
verify-network.sh |
Asserts a node's datapath matches its declared planes — EIP reachability, which plane carries Geneve, and which interfaces services bind. Catches what smoke-test.sh cannot, since none of its traffic leaves the box. |
check-coverage.sh |
Fails the build if total Go coverage falls below the minimum threshold, excluding packages exempt from the requirement. |
diff-coverage.sh |
Coverage check scoped to changed lines only, against an auto-detected base ref (HEAD~1 on main, origin/main on dev, otherwise origin/dev). |
sync-aws-models.sh |
Replaces the embedded AWS Smithy models under internal/awsmodel/models/ with the ten service files from aws/api-models-aws at the given commit, and records that commit in internal/awsmodel/model_source.go. Takes a full 40-character SHA. |
run-bench.sh |
Launches a benchmark against running Ubuntu instances, auto-detecting the AMI for the host architecture. Tracks nbdkit and perf alongside the run. |
disk-performance.sh |
Runs a fio random 70/30 read-write benchmark swept across 4k/16k/128k/1M. Writes JSON results to /tmp/spinifex-disk-bench; override with BENCH_DIR, OUT_DIR, SIZE, JOBS, BLOCK_SIZES. |
network-performance.sh |
iperf throughput from several clients to one server, driven over SSH, writing a summary.txt of Gbit/s per client. --server-ip sets the address clients dial, so pointing it at instance private addresses measures the VPC overlay rather than the external pool. |
Guest images
| Script | What it does |
|---|---|
build-system-image.sh |
Builds a minimal system AMI (Alpine or Ubuntu) from a manifest. All customisation happens inside the libguestfs appliance, so the build touches no host block device, no host mount, and needs no sudo. --import also registers it as an AMI. |
publish-system-image.sh |
Converts a built raw image to compressed qcow2, writes the .sha256 sidecar the catalog verifier expects, and uploads both to the R2 bucket behind iso.mulgadc.com/system-ami/. |
build-microvm-image.sh |
Builds the minimal Alpine kernel + initramfs used by the QEMU microvm machine type and the lb-agent, carrying haproxy (ALB) and nginx-stream (NLB) data planes. |
build-qboot.sh |
Builds qboot, the firmware the daemon passes to every microvm with -bios, from a pinned upstream commit in a digest-pinned Debian container. Fails unless the output matches the recorded sha256, which is also Debian 13's qboot.rom. Writes qboot.rom and its GPL-2.0 licence beside the microvm kernel, and the matching source tarball, which releases publish, under build/qboot/. make build-microvm-image runs it. |
mkosi-build.sh |
Runs an mkosi system-image build inside the pinned builder container, so a runner needs only Docker and the toolchain cannot drift. --image names an output and expands to the right profile list. |
mkosi-publish.sh |
Publishes an mkosi-built image to R2 from inside the same pinned container — qcow2 conversion, .sha256 sidecar, and upload. |
mkosi-common.sh |
Shared plumbing sourced by both mkosi scripts. Builds the toolchain image once, with the builder account matched to the invoking user so bind-mounted output is not written back as root. |
mkosi-builder.Dockerfile |
The pinned image-build toolchain. Debian-hosted regardless of target distro, since mkosi bootstraps the target from its own archive. |
images/ — per-image build inputs
Each subdirectory holds a manifest.conf, init scripts, and a setup.sh that build-system-image.sh runs inside the libguestfs appliance after packages and files are placed. *_test.sh files are shell unit tests for the helper of the same name.
| Path | What it builds |
|---|---|
ecs-agent/ |
The Alpine spinifex-ecs-node container-instance AMI (containerd + ecs-agent). setup.sh sets OpenRC exec bits, wires the CNI bin path, and applies the serial-console tweaks. |
ecs-agent-gpu/ |
The Ubuntu ECS GPU container-instance AMI. Same CNI and containerd wiring adjusted for Ubuntu's package layout, plus the headless NVIDIA driver and nvidia-container-toolkit. |
eks-node/ |
The unified Alpine EKS node AMI (K3s server or agent, selected at first boot). setup.sh installs the pinned, SHA-verified K3s binary and the role init scripts. |
eks-node-gpu/ |
The Ubuntu EKS GPU worker AMI. Agent role only — GPU nodes never run the control plane — plus the NVIDIA GPU stack. |
rds-postgres/ |
The spinifex-rds-postgres AMI. setup.sh sets exec bits on the OpenRC services and creates the root-only directory cloud-init drops the agent env file and gateway CA into. |
rds-mariadb/ |
The spinifex-rds-mariadb AMI. Same layout as rds-postgres/, plus the /etc/my.cnf.d include drop-in and the conf.d/mariadb dependency that keeps the packaged service off an unbootstrapped datadir. |
ubuntu-gpu-{amd,nvidia}-setup.sh |
Chroot setup for GPU-capable Ubuntu guest images: ROCm CLI tools and AMD firmware, or the DKMS-built headless NVIDIA driver. Guest-side only — host VFIO setup is spx admin gpu setup. |
images/eks-node/ — in-guest helpers
Installed into the EKS node AMI and run at boot or on a timer. All host communication goes through the AWS gateway over SigV4; the VM never speaks NATS directly.
| Script | What it does |
|---|---|
eks-node-role.sh |
First-boot role selector. Reads SPINIFEX_K3S_ROLE from the cloud-init env file and enables the matching services — nothing is enabled at bake time. |
k3s-prestart.sh |
Server prestart: refuses to start on a restore-block marker, derives config.yaml from its skeleton, stages the konnectivity-agent and CoreDNS manifests, and waits for the token-webhook kubeconfig. |
k3s-agent-prestart.sh |
Agent prestart shared by the OpenRC and systemd units. Requires K3S_URL and K3S_TOKEN, then resolves the kubelet provider-id from IMDSv2 for EKS parity. |
k3s-first-boot.sh |
Runs once after the server is healthy: reads the node-token and admin kubeconfig, rewrites the kubeconfig server address to the cluster NLB endpoint, and publishes both to the host. |
konnectivity-server-prestart.sh |
Waits for the K3s CA and admin kubeconfig, mints the konnectivity serving cert, and writes the apiserver replica count for the caller to source. |
mulga-eks-addon-sync.sh |
Pulls staged managed-addon descriptors through the gateway, renders each bundle into the K3s auto-deploy dir, and reports per-addon delivery status back. |
mulga-eks-etcd-snapshot.sh |
Snapshots K3s embedded etcd to predastore. Installed into two crond periodic dirs and picks its tier — frequent or daily — from the dir it was invoked from. |
mulga-eks-k3s-recovery.sh |
Applies a control-plane etcd-quorum recovery directive before k3s starts. Pull-based and a no-op unless the reconciler escalated a wedged control plane. |
mulga-eks-state-report.sh |
Publishes server health (healthz, node count) to the cluster reconciler. The apiserver binds the VPC node-ip, which the host daemon cannot probe directly. |
mulga-eks-provider-id.sh |
Writes the K3s config drop-in supplying the metadata the EBS CSI driver needs from IMDS: provider-id plus the region, zone and instance-type labels. |
mulga-ebs-byid.sh |
Mints the EBS-style /dev/disk/by-id symlink the upstream EBS CSI node plugin resolves, from the virtio-blk serial. Alpine's busybox mdev has no eudev to do it. |
mulga-mgmt-net.sh |
Boot oneshot bringing a multi-NIC system VM's interfaces up by MAC from the QEMU fw_cfg netcfg blob, before cloud-init's network stage. cloud-init cannot reliably pick the right NIC out of two. |
mulga-vpc-mtu.sh |
Boot oneshot pinning the VPC data NIC MTU so flannel sizes its VXLAN overlay to fit the OVN geneve underlay. Otherwise large frames are silently dropped while handshakes pass. |
demo/ — GPU demo
| Script | What it does |
|---|---|
deploy.sh |
Deploys and starts the four-instance GPU demo (dashboard, YOLO, Ollama, qwen3-vl) from the *_IP and VIDEO_PATH env vars. |
dashboard_server.py |
Serves the demo web UI and proxies the streams from the GPU instances. |
yolo_stream.py |
Runs YOLO detection over a video file, serving an MJPEG stream of annotated frames and an SSE stream of qwen3-vl scene descriptions. |