A proxied KVM may be framed by the console that proxies it

frame-ancestors 'none' forbids framing even from the same origin, so /kvm/ paths say
'self' instead. Without this the iframe was refused before it loaded.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

PLAN.md

@@ -276,6 +276,50 @@ something long in one, switch away and back, refresh — it's all still there.
   a dashboard could silently reset the sampling rate for the whole machine, overriding the
   fleet refresh rate set in Settings. The sampler now has one source: `refreshMs` in the UI
   prefs, which the hub fans out, remembered per machine as `sampler_ms`.
+- **ZFS tab, with scheduled snapshots** — **shipped 2026-09-24** (`internal/storage/snapshots.go`,
+  `internal/api/zfs.go`, `ZFS.tsx`). Pools, datasets and snapshots moved out of Disks —
+  Disks is the hardware, ZFS is the filesystem on top of it — and the tab gained what it was
+  split out for: **snapshot schedules**. Hourly, daily, weekly or monthly, per dataset,
+  optionally recursive, with a **retention period** and a floor that keeps the newest few
+  whatever their age. A missed window (the machine was off) runs once on return rather than
+  being skipped.
+  **The naming is zfSnap's, deliberately.** `grid` has run zfSnap v1.11.1 from root cron at
+  02:30 with a 6-week TTL since before this repo existed, and it works. So rather than
+  invent a second convention for two tools to disagree over, this writes and reads exactly
+  what zfSnap does — `tank@2026-09-24_02.30.01--6w`, **retention encoded in the name**. Three
+  consequences: Ham Lab shows and expires the snapshots zfSnap already took; the two can run
+  side by side without fighting, since both only remove what has passed the expiry written
+  into its own name; and changing a schedule never re-dates old snapshots, because each
+  carries the rule it was taken under. The page **detects the existing cron job**
+  (zfSnap, zfs-auto-snapshot, sanoid, or a bare `zfs snapshot`) and says so, warning if a Ham
+  Lab schedule is running as well.
+  **Nothing is destroyed unseen.** Only names that parse as a stamp and a TTL are ever
+  candidates for expiry, so a snapshot taken by hand is invisible to a schedule; Run now
+  says that expired snapshots go with it; and a snapshot deleted by hand is confirmed first
+  and needs its full name echoed back by the caller. (A "what would it delete?" preview
+  shipped and was removed the same day as redundant — Mitchell's call.)
+  The datasets table is a **snapshot explorer**: each dataset shows how many snapshots it has
+  and what they cost, and opens into the snapshots themselves — when each was taken, **when
+  it expires** (read from the TTL in its own name, so no server call), its size, and Delete.
+  Deleting takes the same-named snapshot on descendant datasets too, which is how a recursive
+  schedule made it.
+  **A schedule is a real cron job** (2026-09-24, Mitchell's call after the first cut ran it
+  from a ticker inside the daemon). Ham Lab writes `/etc/cron.d/hamlab-zfs`, owns it, and
+  rewrites it whole on every change; each line calls `hamlab zfs-snapshot <id>` as the
+  account Ham Lab runs as, reaching zfs through sudo as the daemon does. Snapshots are
+  backup-shaped and must not stop because a console is down, upgrading, or uninstalled — and
+  a cron line is greppable by anyone looking at the machine the normal way. **A cron
+  expression is the single description of when a schedule fires**: the four common shapes
+  render to one, `custom` states one outright, and the next run is read back with the same
+  evaluator the Jobs tab uses on everyone else's lines, so the page, the file and the
+  prediction cannot disagree. A run's outcome is written to `zfs-runs/<id>.json` beside the
+  config, never to `state.json`, because the cron job is a separate process and the daemon
+  would clobber it. An existing external job can be **adopted** — its schedule, retention
+  and dataset read out of the cron line into a new Ham Lab schedule.
+  **zfSnap is retired on `grid`** as of that day: Ham Lab now takes the nightly `tank`
+  snapshot, the package is purged and the old cron line is gone, with the 515 snapshots it
+  had taken left untouched and fully readable here — which is what adopting its naming
+  bought. See `hamlet/docs/hosts.md`.
 - **Skins** — 13 since 2026-09-19 (`themes.ts`: a palette is one `mk(…)` line): Dark,
   Light, Cyberpunk, Retro, **HUD** (hairlines of cyan on near-black, corner ticks, thin
   type), Nord, Gruvbox, Catppuccin + Latte, Dracula, Tokyo Night, Solarized Dark + Light;
@@ -465,11 +509,12 @@ something long in one, switch away and back, refresh — it's all still there.
 
 ## To do
 
-- **Deprecate copyparty** — asked for 2026-09-23, not started. `copyparty` on `grid`
+- **Grow the Files tab until copyparty can go** — asked for 2026-09-23 as two items (Files
-  (`copyparty.service`, `:3923`, `/etc/copyparty.conf`, serving all of `/tank` as user
+  polish, and deprecating copyparty), merged 2026-09-24: they are the same tab and the same
-  `core` with one account behind NPM) overlaps the **Files** tab almost entirely. The aim
+  work. `copyparty` on `grid` (`copyparty.service`, `:3923`, `/etc/copyparty.conf`, serving
-  is to stop the service and keep only Ham Lab. What Files must gain first, in the order
+  all of `/tank` as user `core` with one account behind NPM) overlaps **Files** almost
-  that matters:
+  entirely. The aim is to stop the service and keep only Ham Lab, with general ease-of-use
+  polish along the way. What Files must gain first, in the order that matters:
   - **Multi-select and bulk actions.** copyparty selects many rows and moves, deletes or
     downloads them together; Files acts on one row at a time.
   - **Download a folder**, and a multi-file download, as a streamed zip or tar — the thing
@@ -489,54 +534,46 @@ something long in one, switch away and back, refresh — it's all still there.
     `hamlet`'s `docs/hosts.md` bare-metal list, `docs/gotchas.md` (the "`docker ps` is not
     the inventory" note names it) and the `grid` row in `CLAUDE.md`.
 
-- **Deprecate Dockhand** — asked for 2026-09-23, not started; the *homelab* half of this
+- **Grow the Docker tab until Dockhand can go** — asked for 2026-09-23 and widened
-  (migrating every stack off Dockhand's data dir, then removing the container) is
+  2026-09-24 to take in Docker tab polish and feature work, on the same reasoning as the
-  `hamlet`'s `docs/pending.md` → "Retire Dockhand; Ham Lab becomes the orchestrator",
+  Files item: it is one tab, and the gaps that block retiring the old tool and the things
-  agreed 2026-09-20. This entry is the *app* half only: what Ham Lab must do before that
+  that would simply make it better are the same body of work. The *homelab* half —
-  migration can start. Its Stacks page already creates, edits, deploys, updates, restarts,
+  migrating every stack off Dockhand's data dir, then removing the container — is `hamlet`'s
-  stops, downs and deletes stacks on every machine, so the gaps are Dockhand's extras:
+  `docs/pending.md` → "Retire Dockhand; Ham Lab becomes the orchestrator", agreed
-  - **Git-sourced stacks.** Dockhand's `stack_sources` table points a stack at a repo and
+  2026-09-20. This entry is the *app* half.
-    pulls it. Ham Lab has no equivalent. Decide whether to build it or to drop it — most
+  Where it stands: Stacks already creates, edits, deploys, updates, restarts, stops, downs
-    stacks here have no row in that table anyway, so dropping it may be honest.
+  and deletes stacks on every machine, and Containers does start/stop/restart, logs that
-  - **Scheduled image updates.** Ham Lab checks for new images only when asked (a registry
+  follow, inspect, and the images/volumes/networks tables. So this is not a rewrite.
-    request Docker Hub counts) and never updates by itself. If anything relies on Dockhand
+  - **What Dockhand still does that Ham Lab does not.**
-    updating on a schedule, that must move here or be dropped on purpose.
+    - *Git-sourced stacks.* Dockhand's `stack_sources` table points a stack at a repo and
-  - **Webhooks / per-host environments**, if either is in use — inventory before deciding.
+      pulls it. Ham Lab has no equivalent. Decide whether to build it or drop it — most
-  - **Deploying from a host path without an orchestrator container.** Ham Lab already runs
+      stacks here have no row in that table anyway, so dropping it may be the honest answer.
-    compose inside `dockhand` because that is how the stacks were deployed; once they move
+    - *Scheduled image updates.* Ham Lab checks for new images only when asked, since each
-    to a neutral home the "via" path is unused for them, but a stack deployed to `grid` or
+      check is a registry request Docker Hub counts, and never updates by itself. If
-    the Mac still goes over `tcp://<node>:2375` from the hub. Decide whether Ham Lab should
+      anything relies on Dockhand updating on a schedule, that must move here or be dropped
-    instead ask the node's own Ham Lab to deploy locally, which would end the `*:2375`
+      on purpose. Note the Jobs tab and the ZFS schedules already show the shape this would
-    exposure (its own item in `docs/pending.md`).
+      take: a cron line Ham Lab owns, not a ticker in the daemon.
-  - *Not app work, but the trigger:* nothing here blocks starting the migration except the
+    - *Webhooks and per-host environments*, if either is in use. Inventory before deciding.
+  - **Known gaps in the tab itself**, worth closing whether or not Dockhand goes:
+    - *Removing networks* — volumes and images can be pruned and removed, networks cannot.
+    - *Container detail is thin next to the stack view.* No per-container resource limits,
+      restart-policy editing, or exec-into-container from the tab (the Terminal tab is the
+      workaround). Decide which of these belong here at all.
+    - *Logs are polled with `--since` rather than followed over a held connection.* That was
+      deliberate and works, but a long tail is choppier than the Services tab's live tail,
+      which does hold a connection. Worth reconciling the two.
+    - *Image update checks are remembered six hours and asked for per stack*, so there is no
+      one view of "what is out of date across the fleet".
+  - **The `*:2375` question, which this is the moment to settle.** Ham Lab runs compose
+    inside `dockhand` because that is how the stacks were deployed; once they move to a
+    neutral home that path is unused for them, but a stack deployed to `grid` or the Mac
+    still goes over `tcp://<node>:2375` from the hub. Decide whether the hub should instead
+    ask the node's own Ham Lab to deploy locally, which would end the unauthenticated
+    Docker socket exposure — its own item in `docs/pending.md`, currently filed under
+    "decided, no action", and this is the change that would make closing it free.
+  - *Not app work, but the trigger:* nothing above blocks starting the migration except the
     inventory. Do that first.
 
-- **Rename the Space tab** — asked for 2026-09-23, not started. "Space" is vague; the tab
-  is WizTree for a disk or pool. Candidates: **Usage**, **Disk Usage**, **Storage**. Note
-  `Disks` already exists and `Storage` would muddle with it, so Usage is probably the one.
-  The rename touches the rail label, the `View` id, the hash route (`#/<machine>/space`,
-  which people may have bookmarked — decide whether to keep `space` as the id and change
-  only the label), the footer's page name and `Space.tsx`'s own strings.
-
-- **ZFS gets its own tab, under Disks** — asked for 2026-09-23, not started. Today ZFS is a
-  section inside `Disks.tsx`: pools with health, capacity, fragmentation, scrub controls, a
-  dataset tree and per-dataset snapshot lists with "Snapshot now". Moving it out makes room
-  for what is actually wanted: **scheduled snapshots with a retention policy**, configured
-  from the console rather than by hand. What exists to build on — `grid` runs `zfSnap`
-  nightly at 02:30 from root cron with 6-week expiry (`zfsnap-cron-6w-daily.sh`, predates
-  this repo), and Ham Lab's Cron tab can already read, write and run root crontabs, so the
-  job can be managed rather than reinvented. Decide: does Ham Lab drive `zfSnap`/`zfs-auto-
-  snapshot` if present, or write and own its own timer? Per-dataset policies or one for the
-  pool? And it must show what a retention rule *will* delete before it runs.
-
-- **Rename Cron to something generic** — asked for 2026-09-23, not started. Mitchell's
-  reasoning: the Services tab is systemd but "Services" is the right generic name for it,
-  so the scheduling tab should be named for the job it does, not the tool. **Scheduled** or
-  **Jobs**. The tab already covers more than cron — `/etc/crontab`, `/etc/cron.d`, user
-  crontabs, `run-parts` directories *and* systemd timers — so the generic name is the
-  honest one. Same rename surface as the Space tab: rail label, `View` id, hash route,
-  footer name.
-
 - **A Virtualization tab** — asked for 2026-09-23, not started, and **large**. Run and
   manage **LXC containers and virtual machines** from the console, the way Proxmox does —
   which is the comparison Ham Lab has invited since its first line. Nothing exists yet.
@@ -548,10 +585,39 @@ something long in one, switch away and back, refresh — it's all still there.
   `internal/term`); snapshots and clones; and what it does on a host with no virtualisation
   at all, which is every host here today. Worth scoping as its own plan before starting.
 
-- **Files explorer: cleanup and ease of use** — asked for 2026-09-23, not started.
+- **A KVM tab** — asked for 2026-09-24, not started. **Dead simple is the brief**: paste one
-  General polish and ease of use. The **Deprecate copyparty**
+  HTTP link, the tab shows whatever UI that KVM serves, and the keyboard goes to it. **This
-  item names the heavier pieces (multi-select and bulk actions, folder download, resumable
+  is KVM the keyboard-video-mouse switch, not KVM the hypervisor** — the Virtualization item
-  uploads, a gallery, search) — worth doing the two together, since they are the same tab.
+  above means the other one.
+  - **The device**: a **Sipeed NanoKVM** attached to `grid` at **`10.0.0.217`** (MAC
+    `48:da:35:6f:d0:31`), web UI on port 80. The only one so far, so store a list, not one URL.
+  - **Framing works.** Probed 2026-09-24: it sends **no `X-Frame-Options` and no
+    `frame-ancestors`**, so it does not refuse to be embedded. No fallback needed.
+  - **Proxy it through the daemon — that is what makes the simple version possible.** The
+    KVM is **HTTP only**, and Ham Lab is served over TLS, so a bare `http://…` iframe is
+    blocked as mixed content and the "just paste a link" idea dies there. Serving it from Ham
+    Lab's own origin (`/api/kvm/<id>/…`, on the `proxyNode` machinery, which already carries
+    WebSockets — exactly what the video and input streams are) fixes three things at once:
+    no mixed content, the KVM sits behind **Ham Lab's** login rather than a second one
+    (`/api/vm/info` answers 401, so it has its own, which the hub can hold once), and the
+    iframe becomes **same-origin**, which matters for the keyboard below.
+  - **The keyboard needs less than it sounds.** A parent page *cannot* capture keys and
+    forward them into a **cross-origin** iframe — the browser forbids synthetic events across
+    origins — but it does not need to: an iframe with focus receives keystrokes directly, and
+    the NanoKVM's own page already handles them. So "capture and give it to the iframe" is
+    really: click-to-focus, keep focus there, and stop the console's own shortcuts from
+    eating keys while it is focused. What is genuinely worth adding is **fullscreen** plus
+    `navigator.keyboard.lock()`, which is what stops the *browser* swallowing Escape, Tab and
+    Ctrl-W so they reach the guest — note it is Chromium-only, so Firefox and iOS Safari get
+    the plain version. Proxying makes the iframe same-origin anyway, so real forwarding stays
+    available if some KVM ever needs it.
+  - **Where the list lives is the one real design decision.** A KVM earns its keep when the
+    machine is *down* — no OS, a bad kernel, a BIOS screen — which is when that host's Ham Lab
+    cannot answer. So it is **hub-side** and must work for machines running no Ham Lab at all.
+    A `kvm_url` in each node's own `config.json` would fail in exactly the case it is for.
+  - **Do not give it an NPM host** without deciding to: `*.bigham.dev` is a wildcard to the
+    public IP, so that would put BIOS-level control of `grid` on the internet behind only the
+    KVM's own login. Behind Ham Lab's proxy it inherits Ham Lab's auth instead.
 
 ## Open questions
 

cmd/hamlab/main.go

@@ -4,6 +4,7 @@ package main
 import (
 	"bufio"
 	"context"
+	"encoding/json"
 	"errors"
 	"flag"
 	"fmt"
@@ -32,6 +33,7 @@ import (
 	"hamlab/internal/metrics"
 	"hamlab/internal/nodes"
 	"hamlab/internal/services"
+	"hamlab/internal/priv"
 	"hamlab/internal/setup"
 	"hamlab/internal/space"
 	"hamlab/internal/storage"
@@ -59,6 +61,7 @@ func main() {
   serve                run the console in the foreground (what the service runs)
   passwd [user]        set a user's password (makes the account if it is new; reads stdin when piped)
   users                list the accounts
+  zfs-snapshot <id>    run one ZFS snapshot schedule (what its cron line calls)
   user-rm <user>       remove an account and sign it out everywhere
   check                verify the config and state load, then exit
   version
@@ -129,6 +132,10 @@ func main() {
 		if err := passwd(cfg, flag.Arg(1)); err != nil {
 			log.Fatal(err)
 		}
+	case "zfs-snapshot":
+		if err := zfsSnapshot(cfg, flag.Arg(1)); err != nil {
+			log.Fatal(err)
+		}
 	case "users":
 		migrateUsers(cfg)
 		if len(cfg.Users) == 0 {
@@ -178,6 +185,58 @@ func main() {
 
 // migrateUsers gives a pre-users config its first account, named after this OS user, so
 // the CLI sees the same accounts the daemon will.
+// zfsSnapshot runs one snapshot schedule: take it, then expire what has aged out. This is
+// what the cron line Ham Lab writes actually calls, so a schedule keeps running whether or
+// not the console is up — and keeps running if Ham Lab is never opened again.
+//
+// It deliberately does not write state.json. The daemon owns that file and would clobber
+// anything written from out here; the outcome goes beside it instead, in zfs-runs/.
+func zfsSnapshot(cfg *config.Config, id string) error {
+	if strings.TrimSpace(id) == "" {
+		return errors.New("usage: hamlab zfs-snapshot <schedule-id>")
+	}
+	st, err := store.Open(filepath.Join(cfg.DataDir(), "state.json"))
+	if err != nil {
+		return err
+	}
+	var raw json.RawMessage
+	st.View(func(s *store.State) { raw = s.ZFS })
+	var z struct {
+		Policies []storage.Policy `json:"policies"`
+	}
+	if len(raw) > 0 {
+		if err := json.Unmarshal(raw, &z); err != nil {
+			return fmt.Errorf("reading the schedules: %w", err)
+		}
+	}
+	var p *storage.Policy
+	for i := range z.Policies {
+		if z.Policies[i].ID == id {
+			p = &z.Policies[i]
+		}
+	}
+	if p == nil {
+		return fmt.Errorf("no snapshot schedule %q — it may have been removed; this cron line is stale", id)
+	}
+	res, runErr := storage.New(priv.Available()).Run(context.Background(), p)
+	rec := storage.RunRecord{At: time.Now(), Taken: res.Taken, Pruned: res.Pruned}
+	if runErr != nil {
+		rec.Error = runErr.Error()
+	}
+	if err := storage.WriteRun(cfg.DataDir(), id, rec); err != nil {
+		fmt.Fprintln(os.Stderr, "recording the run:", err)
+	}
+	if runErr != nil {
+		return runErr
+	}
+	fmt.Printf("%s: took %s", p.Name, res.Taken)
+	if len(res.Pruned) > 0 {
+		fmt.Printf(", expired %d", len(res.Pruned))
+	}
+	fmt.Println()
+	return nil
+}
+
 func migrateUsers(cfg *config.Config) {
 	name := os.Getenv("USER")
 	if u, err := user.Current(); err == nil && u.Username != "" {

internal/api/kvm.go

@@ -0,0 +1,272 @@
+package api
+
+// IP-KVMs: a device's own web page, shown inside Ham Lab.
+//
+// The whole difficulty is that Ham Lab is served over TLS and a KVM usually is not, so a
+// browser refuses to put the device's page in an iframe — mixed content, blocked before
+// anything loads. The fix is to serve the device from Ham Lab's own origin, which also
+// means it inherits Ham Lab's login instead of needing a second one, and makes the iframe
+// same-origin.
+//
+// Proxying an arbitrary web app under a path prefix is the hard part, and nothing here
+// knows anything about any particular device. A device UI asks for its assets by absolute
+// path — "/assets/app.js", "/api/whatever" — which a prefix cannot rewrite away. So the
+// prefix is only the entry point: any request Ham Lab does not serve itself is routed to
+// whichever KVM the asking page belongs to, worked out from the Referer, and from a cookie
+// when there is no Referer (a WebSocket upgrade, say). Ham Lab's own routes always win, so
+// a device can never shadow the console.
+
+import (
+	"crypto/rand"
+	"encoding/hex"
+	"encoding/json"
+	"errors"
+	"io"
+	"log"
+	"net/http"
+	"net/http/httputil"
+	"net/url"
+	"strings"
+	"sync"
+	"time"
+
+	"hamlab/internal/config"
+)
+
+const kvmPrefix = "/kvm/"
+const kvmCookie = "hamlab_kvm"
+
+type kvmProxies struct {
+	mu sync.Mutex
+	m  map[string]*httputil.ReverseProxy
+}
+
+func (s *Server) kvm(id string) (config.KVM, bool) {
+	var found config.KVM
+	ok := false
+	s.cfg.View(func(c *config.Config) {
+		for _, k := range c.KVMs {
+			if k.ID == id {
+				found, ok = k, true
+			}
+		}
+	})
+	return found, ok
+}
+
+// kvmProxy is the reverse proxy for one device, made once and kept.
+func (s *Server) kvmProxy(k config.KVM) (*httputil.ReverseProxy, error) {
+	s.kvms.mu.Lock()
+	defer s.kvms.mu.Unlock()
+	if s.kvms.m == nil {
+		s.kvms.m = map[string]*httputil.ReverseProxy{}
+	}
+	if p, ok := s.kvms.m[k.ID]; ok {
+		return p, nil
+	}
+	target, err := url.Parse(k.URL)
+	if err != nil || target.Host == "" {
+		return nil, errors.New("this KVM's address is not a URL")
+	}
+	prefix := kvmPrefix + k.ID
+	p := &httputil.ReverseProxy{
+		Director: func(r *http.Request) {
+			r.URL.Scheme, r.URL.Host = target.Scheme, target.Host
+			// Entry through the prefix keeps the path after it; anything else arrived by its
+			// own absolute path and is passed through untouched.
+			if rest, ok := strings.CutPrefix(r.URL.Path, prefix); ok {
+				if rest == "" {
+					rest = "/"
+				}
+				r.URL.Path = rest
+			}
+			r.URL.RawPath = ""
+			r.Host = target.Host
+			// The device gets its own host and none of Ham Lab's session.
+			r.Header.Del("Origin")
+			stripCookie(r, cookieName, kvmCookie)
+		},
+		FlushInterval: -1, // a video stream must not be buffered
+		ModifyResponse: func(resp *http.Response) error {
+			// The device's own framing and transport policy would stop it being embedded, or
+			// send a browser off to plain HTTP. Ham Lab is the origin now, and says what it
+			// serves; its own headers are added on the way out.
+			for _, h := range []string{"X-Frame-Options", "Content-Security-Policy", "Content-Security-Policy-Report-Only", "Strict-Transport-Security", "Referrer-Policy"} {
+				resp.Header.Del(h)
+			}
+			// A redirect to the device's own root has to come back through the prefix.
+			if loc := resp.Header.Get("Location"); loc != "" {
+				if u, err := url.Parse(loc); err == nil && (u.Host == "" || u.Host == target.Host) && strings.HasPrefix(u.Path, "/") && !strings.HasPrefix(u.Path, prefix) {
+					u.Scheme, u.Host = "", ""
+					u.Path = prefix + u.Path
+					resp.Header.Set("Location", u.String())
+				}
+			}
+			return nil
+		},
+		ErrorHandler: func(w http.ResponseWriter, r *http.Request, err error) {
+			w.Header().Set("Content-Type", "application/json")
+			w.WriteHeader(http.StatusBadGateway)
+			_ = json.NewEncoder(w).Encode(map[string]string{"error": k.Name + " unreachable: " + err.Error()})
+		},
+	}
+	s.kvms.m[k.ID] = p
+	return p, nil
+}
+
+// stripCookie removes named cookies from a request, so Ham Lab's session is never handed to
+// a device and the device never sees which KVM the browser is looking at.
+func stripCookie(r *http.Request, names ...string) {
+	cs := r.Cookies()
+	kept := cs[:0]
+	for _, c := range cs {
+		drop := false
+		for _, n := range names {
+			if c.Name == n {
+				drop = true
+			}
+		}
+		if !drop {
+			kept = append(kept, c)
+		}
+	}
+	r.Header.Del("Cookie")
+	for _, c := range kept {
+		r.AddCookie(c)
+	}
+}
+
+func (s *Server) forgetKVMProxy(id string) {
+	s.kvms.mu.Lock()
+	delete(s.kvms.m, id)
+	s.kvms.mu.Unlock()
+}
+
+// serveKVM is the entry point: /kvm/<id>/… . It also drops a cookie naming the device, so
+// that requests the page makes later by absolute path can be traced back to it.
+func (s *Server) serveKVM(w http.ResponseWriter, r *http.Request) {
+	id := r.PathValue("id")
+	k, ok := s.kvm(id)
+	if !ok {
+		jsonErr(w, 404, "no such KVM")
+		return
+	}
+	p, err := s.kvmProxy(k)
+	if err != nil {
+		jsonErr(w, 400, err.Error())
+		return
+	}
+	http.SetCookie(w, &http.Cookie{Name: kvmCookie, Value: id, Path: "/", HttpOnly: true, Secure: secure(r), SameSite: http.SameSiteLaxMode})
+	p.ServeHTTP(w, r)
+}
+
+// kvmFor works out which device a request belongs to when it arrived by absolute path
+// rather than through the prefix: the Referer names the page that asked, and a cookie
+// covers what has no Referer at all. Nothing device-specific: it never has to know which
+// paths a given KVM uses.
+func (s *Server) kvmFor(r *http.Request) (config.KVM, bool) {
+	if ref := r.Header.Get("Referer"); ref != "" {
+		if u, err := url.Parse(ref); err == nil {
+			if rest, ok := strings.CutPrefix(u.Path, kvmPrefix); ok {
+				if id, _, _ := strings.Cut(rest, "/"); id != "" {
+					if k, ok := s.kvm(id); ok {
+						return k, true
+					}
+				}
+			}
+		}
+		// asked for by a page that is not a KVM: not ours to answer
+		return config.KVM{}, false
+	}
+	if c, err := r.Cookie(kvmCookie); err == nil && c.Value != "" {
+		return s.kvm(c.Value)
+	}
+	return config.KVM{}, false
+}
+
+// kvmFallback answers a request Ham Lab does not serve itself, when it came from inside a
+// KVM's page. It returns false when the request has nothing to do with a KVM, and the
+// caller carries on with its own 404.
+func (s *Server) kvmFallback(w http.ResponseWriter, r *http.Request) bool {
+	k, ok := s.kvmFor(r)
+	if !ok {
+		return false
+	}
+	p, err := s.kvmProxy(k)
+	if err != nil {
+		return false
+	}
+	p.ServeHTTP(w, r)
+	return true
+}
+
+// ---- the list ----
+
+func (s *Server) listKVMs(w http.ResponseWriter, _ *http.Request) {
+	out := []config.KVM{}
+	s.cfg.View(func(c *config.Config) { out = append(out, c.KVMs...) })
+	jsonOK(w, 200, out)
+}
+
+func (s *Server) addKVM(w http.ResponseWriter, r *http.Request) {
+	var body struct{ Name, URL, Machine string }
+	if err := json.NewDecoder(io.LimitReader(r.Body, 4096)).Decode(&body); err != nil {
+		jsonErr(w, 400, "bad request")
+		return
+	}
+	raw := strings.TrimSpace(body.URL)
+	if raw != "" && !strings.Contains(raw, "://") {
+		raw = "http://" + raw // what you would type: an address, not a scheme
+	}
+	u, err := url.Parse(raw)
+	if err != nil || (u.Scheme != "http" && u.Scheme != "https") || u.Host == "" {
+		jsonErr(w, 400, "address: http(s)://host, or just the host")
+		return
+	}
+	name := strings.TrimSpace(body.Name)
+	if name == "" {
+		name = u.Host
+	}
+	if len(name) > maxNameLen {
+		jsonErr(w, 400, "that name is too long")
+		return
+	}
+	b := make([]byte, 6)
+	if _, err := rand.Read(b); err != nil {
+		jsonErr(w, 500, err.Error())
+		return
+	}
+	k := config.KVM{ID: hex.EncodeToString(b), Name: name, URL: u.Scheme + "://" + u.Host + strings.TrimSuffix(u.Path, "/"), Machine: strings.TrimSpace(body.Machine), Added: time.Now()}
+	if err := s.cfg.Update(func(c *config.Config) { c.KVMs = append(c.KVMs, k) }); err != nil {
+		jsonErr(w, 500, err.Error())
+		return
+	}
+	log.Printf("kvm %q added (%s) by %s", k.Name, k.URL, clientIP(r))
+	jsonOK(w, 201, k)
+}
+
+func (s *Server) removeKVM(w http.ResponseWriter, r *http.Request) {
+	id := r.PathValue("id")
+	found := false
+	if err := s.cfg.Update(func(c *config.Config) {
+		kept := c.KVMs[:0]
+		for _, k := range c.KVMs {
+			if k.ID == id {
+				found = true
+				continue
+			}
+			kept = append(kept, k)
+		}
+		c.KVMs = kept
+	}); err != nil {
+		jsonErr(w, 500, err.Error())
+		return
+	}
+	if !found {
+		jsonErr(w, 404, "no such KVM")
+		return
+	}
+	s.forgetKVMProxy(id)
+	log.Printf("kvm %s removed by %s", id, clientIP(r))
+	w.WriteHeader(204)
+}

internal/api/server.go

@@ -69,6 +69,7 @@ type Server struct {
 	storage     *storage.Manager
 	space       *space.Manager
 	update      updateState
+	kvms        kvmProxies
 	fsLocal     files.FS
 	fsSudo      files.FS // nil without passwordless sudo
 	nodes       *nodes.Manager
@@ -251,6 +252,11 @@ func New(cfg *config.Config, sessions *auth.Sessions, mgr *term.Manager, col *me
 	m.HandleFunc("GET /api/storage/zfs/snapshots", s.auth(s.zfsSnapshots))
 	m.HandleFunc("POST /api/storage/zfs/scrub/{pool}/{action}", s.auth(s.zfsScrub))
 	m.HandleFunc("POST /api/storage/zfs/snapshot", s.auth(s.zfsSnapshot))
+	m.HandleFunc("GET /api/zfs", s.auth(s.zfsOverview))
+	m.HandleFunc("PUT /api/zfs/schedules", s.auth(s.zfsPolicyPut))
+	m.HandleFunc("DELETE /api/zfs/schedules/{id}", s.auth(s.zfsPolicyDelete))
+	m.HandleFunc("POST /api/zfs/schedules/{id}/run", s.auth(s.zfsRunNow))
+	m.HandleFunc("POST /api/zfs/snapshots/delete", s.auth(s.zfsSnapshotDelete))
 	m.HandleFunc("GET /api/space", s.auth(s.spaceOverview))
 	m.HandleFunc("GET /api/space/{disk}", s.auth(s.spaceStatus))
 	m.HandleFunc("POST /api/space/{disk}/scan", s.auth(s.spaceScan))
@@ -279,7 +285,20 @@ func New(cfg *config.Config, sessions *auth.Sessions, mgr *term.Manager, col *me
 	m.HandleFunc("PUT /api/prefs/sampler", s.auth(s.putSampler))
 	m.HandleFunc("GET /api/prefs/history", s.auth(s.getHistoryWindow))
 	m.HandleFunc("PUT /api/prefs/history", s.auth(s.putHistoryWindow))
-	m.HandleFunc("/api/", func(w http.ResponseWriter, _ *http.Request) { jsonErr(w, 404, "not found") })
+	m.HandleFunc("GET /api/kvms", s.auth(s.listKVMs))
+	m.HandleFunc("POST /api/kvms", s.auth(s.addKVM))
+	m.HandleFunc("DELETE /api/kvms/{id}", s.auth(s.removeKVM))
+	m.HandleFunc("/kvm/{id}/{rest...}", s.auth(s.serveKVM))
+	m.HandleFunc("/kvm/{id}", s.auth(s.serveKVM))
+	// A path Ham Lab does not serve may still belong to a KVM's page, which asks for its
+	// own assets by absolute path. Ham Lab's own routes matched first, so this can only
+	// catch what would otherwise have been a 404.
+	m.HandleFunc("/api/", s.auth(func(w http.ResponseWriter, r *http.Request) {
+		if s.kvmFallback(w, r) {
+			return
+		}
+		jsonErr(w, 404, "not found")
+	}))
 	m.HandleFunc("/", s.static)
 	s.mux = m
 	return s
@@ -302,9 +321,16 @@ func sensorPath(p string) bool {
 func (s *Server) Handler() http.Handler {
 	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
 		h := w.Header()
-		// The console must never be framed. An embedded widget exists to be: it is a page
+		// The console must never be framed. Two exceptions: an embedded widget exists to be
-		// that can only read sensors, and only with a read-only token in its URL.
+		// (a page that can only read sensors, and only with a read-only token in its URL),
-		if r.URL.Path != "/embed/widget" {
+		// and a proxied KVM is framed by the console itself — 'none' forbids that even from
+		// the same origin, so those paths say 'self' instead of opening up.
+		switch {
+		case r.URL.Path == "/embed/widget":
+		case strings.HasPrefix(r.URL.Path, kvmPrefix):
+			h.Set("X-Frame-Options", "SAMEORIGIN")
+			h.Set("Content-Security-Policy", "frame-ancestors 'self'")
+		default:
 			h.Set("X-Frame-Options", "DENY")
 			h.Set("Content-Security-Policy", "frame-ancestors 'none'")
 		}
@@ -350,6 +376,17 @@ func (s *Server) setCookie(w http.ResponseWriter, r *http.Request, tok string, m
 	})
 }
 
+// signedIn reports whether a request carries a live browser session, without the
+// side effects of the auth middleware.
+func (s *Server) signedIn(r *http.Request) bool {
+	c, err := r.Cookie(cookieName)
+	if err != nil {
+		return false
+	}
+	_, ok := s.sessions.Lookup(c.Value, clientKey(r))
+	return ok
+}
+
 // auth accepts the browser session cookie or a bearer API token issued to a hub.
 func (s *Server) auth(next http.HandlerFunc) http.HandlerFunc {
 	return func(w http.ResponseWriter, r *http.Request) {
@@ -2440,6 +2477,12 @@ func (s *Server) static(w http.ResponseWriter, r *http.Request) {
 			return
 		}
 	}
+	// A path with no file of ours behind it may belong to a KVM's page, which asks for its
+	// own assets by absolute path. Checked here rather than in the router because this
+	// handler is the catch-all — and gated on a session, since this handler is public.
+	if s.signedIn(r) && s.kvmFallback(w, r) {
+		return
+	}
 	// SPA fallback
 	w.Header().Set("Cache-Control", "no-cache")
 	http.ServeFileFS(w, r, s.ui, "index.html")

internal/api/zfs.go

@@ -0,0 +1,262 @@
+package api
+
+// The ZFS tab: pools, datasets, snapshots, and a schedule for taking and expiring them.
+//
+// grid already runs zfSnap from root cron and has done for longer than this repo has
+// existed, so this does not replace it. It reads the same snapshot names, expires by the
+// same rule, and surfaces the cron job that drives it — so what is on screen is the truth
+// about the machine, whether Ham Lab or zfSnap is doing the work. A Ham Lab schedule is for
+// a machine that has no such tool, and the page says plainly when both are running.
+
+import (
+	"context"
+	"crypto/rand"
+	"encoding/hex"
+	"encoding/json"
+	"io"
+	"log"
+	"net/http"
+	"os"
+	"path/filepath"
+	"strings"
+	"time"
+
+	"hamlab/internal/storage"
+	"hamlab/internal/store"
+)
+
+type zfsState struct {
+	Policies []storage.Policy `json:"policies"`
+}
+
+func (s *Server) zfsLoad() zfsState {
+	var raw json.RawMessage
+	s.store.View(func(st *store.State) { raw = st.ZFS })
+	var z zfsState
+	if len(raw) > 0 {
+		_ = json.Unmarshal(raw, &z)
+	}
+	if z.Policies == nil {
+		z.Policies = []storage.Policy{}
+	}
+	return z
+}
+
+func (s *Server) zfsSave(z zfsState) error {
+	b, err := json.Marshal(z)
+	if err != nil {
+		return err
+	}
+	return s.store.Update(func(st *store.State) { st.ZFS = b })
+}
+
+// externalSnapshotJob finds a cron job that is already taking snapshots on this machine, so
+// the page can say so rather than let two schedules quietly double up.
+func (s *Server) externalSnapshotJob(ctx context.Context) map[string]any {
+	if s.cron == nil {
+		return nil
+	}
+	ov := s.cron.Overview(ctx)
+	for _, j := range ov.Jobs {
+		if !j.Enabled {
+			continue
+		}
+		c := strings.ToLower(j.Command)
+		if !strings.Contains(c, "zfsnap") && !strings.Contains(c, "zfs-auto-snapshot") && !strings.Contains(c, "sanoid") &&
+			!(strings.Contains(c, "zfs ") && strings.Contains(c, "snapshot")) {
+			continue
+		}
+		return map[string]any{
+			"tool": toolName(c), "schedule": j.Schedule, "command": j.Command,
+			"user": j.User, "source": j.Source, "id": j.ID, "next": j.Next, "last": j.Last,
+		}
+	}
+	return nil
+}
+
+func toolName(cmd string) string {
+	switch {
+	case strings.Contains(cmd, "zfsnap"):
+		return "zfSnap"
+	case strings.Contains(cmd, "zfs-auto-snapshot"):
+		return "zfs-auto-snapshot"
+	case strings.Contains(cmd, "sanoid"):
+		return "sanoid"
+	}
+	return "a cron job"
+}
+
+// randomID is a short opaque id for a schedule.
+func randomID() (string, error) {
+	b := make([]byte, 6)
+	if _, err := rand.Read(b); err != nil {
+		return "", err
+	}
+	return hex.EncodeToString(b), nil
+}
+
+// ---- the cron file ----
+//
+// A schedule is a cron line, written to /etc/cron.d/hamlab-zfs, which Ham Lab owns and
+// rewrites whole whenever the schedules change. Nothing in the daemon runs them: snapshots
+// are backup-shaped and should not stop because a web console is down, being upgraded, or
+// uninstalled. The line calls `hamlab zfs-snapshot <id>` as the account Ham Lab runs as,
+// which reaches zfs through sudo exactly as the daemon does.
+
+// syncCron puts the current schedules on the machine.
+func (s *Server) syncCron(ctx context.Context, z zfsState) error {
+	exe, err := os.Executable()
+	if err != nil {
+		return err
+	}
+	exe, _ = filepath.EvalSymlinks(exe)
+	return s.storage.WriteCronFile(ctx, z.Policies, s.user, exe, s.cfg.Path())
+}
+
+// ---- endpoints ----
+
+func (s *Server) zfsOverview(w http.ResponseWriter, r *http.Request) {
+	z := s.zfsLoad()
+	now := time.Now()
+	dir := s.cfg.DataDir()
+	out := make([]map[string]any, 0, len(z.Policies))
+	for i := range z.Policies {
+		p := z.Policies[i]
+		expr, _ := p.CronExpr()
+		m := map[string]any{"policy": p, "cron": expr, "last": storage.ReadRun(dir, p.ID)}
+		if p.Enabled {
+			if n := p.NextRun(now); !n.IsZero() {
+				m["next"] = n
+			}
+		}
+		out = append(out, m)
+	}
+	jsonOK(w, 200, map[string]any{
+		"zfs": s.storage.ZFS(r.Context()), "policies": out, "external": s.externalSnapshotJob(r.Context()),
+	})
+}
+
+func (s *Server) zfsPolicyPut(w http.ResponseWriter, r *http.Request) {
+	var p storage.Policy
+	if err := json.NewDecoder(io.LimitReader(r.Body, 16<<10)).Decode(&p); err != nil {
+		jsonErr(w, 400, "bad request")
+		return
+	}
+	if err := p.Validate(); err != nil {
+		jsonErr(w, 400, err.Error())
+		return
+	}
+	z := s.zfsLoad()
+	found := false
+	for i := range z.Policies {
+		if z.Policies[i].ID == p.ID && p.ID != "" {
+			z.Policies[i] = p
+			found = true
+		}
+	}
+	if !found {
+		if p.ID == "" {
+			id, err := randomID()
+			if err != nil {
+				jsonErr(w, 500, err.Error())
+				return
+			}
+			p.ID = id
+		}
+		z.Policies = append(z.Policies, p)
+	}
+	if err := s.zfsSave(z); err != nil {
+		jsonErr(w, 500, err.Error())
+		return
+	}
+	if err := s.syncCron(r.Context(), z); err != nil {
+		jsonErr(w, 500, "saved, but writing "+storage.CronFile+" failed: "+err.Error())
+		return
+	}
+	expr, _ := p.CronExpr()
+	log.Printf("zfs schedule %q saved by %s (%s, %q, keep %s)", p.Name, clientIP(r), p.Dataset, expr, p.KeepFor)
+	jsonOK(w, 200, p)
+}
+
+func (s *Server) zfsPolicyDelete(w http.ResponseWriter, r *http.Request) {
+	id := r.PathValue("id")
+	z := s.zfsLoad()
+	kept := z.Policies[:0]
+	found := false
+	for _, p := range z.Policies {
+		if p.ID == id {
+			found = true
+			continue
+		}
+		kept = append(kept, p)
+	}
+	if !found {
+		jsonErr(w, 404, "no such schedule")
+		return
+	}
+	z.Policies = kept
+	if err := s.zfsSave(z); err != nil {
+		jsonErr(w, 500, err.Error())
+		return
+	}
+	if err := s.syncCron(r.Context(), z); err != nil {
+		jsonErr(w, 500, "removed, but writing "+storage.CronFile+" failed: "+err.Error())
+		return
+	}
+	storage.ForgetRun(s.cfg.DataDir(), id)
+	log.Printf("zfs schedule %s removed by %s (its snapshots are left alone)", id, clientIP(r))
+	w.WriteHeader(204)
+}
+
+func (s *Server) zfsPolicy(id string) (storage.Policy, bool) {
+	for _, p := range s.zfsLoad().Policies {
+		if p.ID == id {
+			return p, true
+		}
+	}
+	return storage.Policy{}, false
+}
+
+func (s *Server) zfsRunNow(w http.ResponseWriter, r *http.Request) {
+	id := r.PathValue("id")
+	p, ok := s.zfsPolicy(id)
+	if !ok {
+		jsonErr(w, 404, "no such schedule")
+		return
+	}
+	res, err := s.storage.Run(r.Context(), &p)
+	rec := storage.RunRecord{At: time.Now(), Taken: res.Taken, Pruned: res.Pruned}
+	if err != nil {
+		rec.Error = err.Error()
+	}
+	_ = storage.WriteRun(s.cfg.DataDir(), id, rec)
+	if err != nil {
+		jsonErr(w, 500, err.Error())
+		return
+	}
+	log.Printf("zfs schedule %q run by %s: took %s, expired %d", p.Name, clientIP(r), res.Taken, len(res.Pruned))
+	jsonOK(w, 200, res)
+}
+
+// zfsSnapshotDelete removes one snapshot, named in full and typed back by the caller.
+func (s *Server) zfsSnapshotDelete(w http.ResponseWriter, r *http.Request) {
+	var body struct {
+		Name      string `json:"name"`
+		Confirm   string `json:"confirm"`
+		Recursive bool   `json:"recursive"`
+	}
+	if err := json.NewDecoder(io.LimitReader(r.Body, 4096)).Decode(&body); err != nil {
+		jsonErr(w, 400, "bad request")
+		return
+	}
+	if body.Confirm != body.Name {
+		jsonErr(w, 400, "type the snapshot's full name to confirm")
+		return
+	}
+	if err := s.storage.DestroySnapshot(r.Context(), body.Name, body.Recursive); err != nil {
+		jsonErr(w, 500, err.Error())
+		return
+	}
+	log.Printf("zfs snapshot %s destroyed by %s%s", body.Name, clientIP(r), map[bool]string{true: " (recursive)", false: ""}[body.Recursive])
+	w.WriteHeader(204)
+}

internal/config/config.go

@@ -61,6 +61,21 @@ type UpdateSource struct {
 	Check string `json:"check,omitempty"`
 }
 
+// KVM is an IP-KVM: a device with a web page that gives screen, keyboard and mouse access
+// to a machine, including when that machine has no OS running. Ham Lab holds the list on
+// the hub and proxies each one, because a KVM matters exactly when the machine it is
+// attached to cannot answer for itself.
+type KVM struct {
+	ID   string `json:"id"`
+	Name string `json:"name"`
+	// URL is the device's own address, as you would type it into a browser on the LAN.
+	URL string `json:"url"`
+	// Machine is which host it is attached to, for display only — it need not be a machine
+	// Ham Lab manages, and usually the point is that it is not answering.
+	Machine string    `json:"machine,omitempty"`
+	Added   time.Time `json:"added"`
+}
+
 type Config struct {
 	Listen string `json:"listen"`
 	// PasswordHash is the single login of installs before users existed. On load it becomes
@@ -70,6 +85,7 @@ type Config struct {
 	TLS          string  `json:"tls,omitempty"` // "auto" → self-signed cert in the data dir; "" → plain HTTP
 	Tokens       []Token `json:"tokens,omitempty"`
 	Nodes        []Node  `json:"nodes,omitempty"`
+	KVMs         []KVM   `json:"kvms,omitempty"`
 	// Updates: where a newer build comes from. See UpdateSource.
 	Updates UpdateSource `json:"update,omitempty"`
 

internal/cron/cron.go

@@ -1,3 +1,6 @@
+// Package cron reads and writes what the UI calls the **Jobs** tab (named "Cron" until
+// 2026-09-24; the package and its /api/cron endpoints kept the old name).
+//
 // Package cron reads every scheduled job on the host into one list: /etc/crontab,
 // /etc/cron.d/*, per-user crontabs, the run-parts directories they drive, and systemd
 // timers. It computes next runs from the cron expressions, pulls last runs from the

internal/cron/schedule.go

@@ -132,6 +132,17 @@ func (s *schedule) dayMatches(t time.Time) bool {
 	}
 }
 
+// NextRun is the first time strictly after `from` that a cron expression fires. Exported so
+// anything that writes a cron line — the ZFS snapshot schedules — can say when it runs next
+// using the same evaluator that reads everyone else's.
+func NextRun(expr string, from time.Time) (time.Time, bool) { return nextRun(expr, from) }
+
+// ValidSchedule reports whether a cron expression parses.
+func ValidSchedule(expr string) error {
+	_, err := parseSchedule(expr)
+	return err
+}
+
 // nextRun finds the first time strictly after `from` that the schedule fires.
 func nextRun(expr string, from time.Time) (time.Time, bool) {
 	s, err := parseSchedule(expr)

internal/metrics/metrics.go

@@ -94,6 +94,11 @@ type FS struct {
 	// sshfs, and so on. Its bytes are somebody else's disk, so a panel about this machine's
 	// hardware leaves it out, while the sensor stays for anyone who does want it.
 	Remote bool `json:"remote,omitempty"`
+	// Disk is the physical drive these bytes are on ("nvme0", "sda"), named as the sampler
+	// names it; empty when that cannot be
+	// said honestly: a ZFS pool or an md array spreads a filesystem over several drives, and
+	// a remote mount is on none of them.
+	Disk string `json:"disk,omitempty"`
 }
 
 type UPS struct {
@@ -825,7 +830,7 @@ func filesystems() []FS {
 				return
 			}
 			bs := uint64(st.Bsize)
-			done <- FS{Mount: mount, Type: fs[2], Size: st.Blocks * bs, Used: (st.Blocks - st.Bfree) * bs, Remote: remoteFS[fs[2]]}
+			done <- FS{Mount: mount, Type: fs[2], Size: st.Blocks * bs, Used: (st.Blocks - st.Bfree) * bs, Remote: remoteFS[fs[2]], Disk: diskOf(fs[0])}
 		}()
 		select {
 		case r := <-done:
@@ -839,6 +844,46 @@ func filesystems() []FS {
 	return out
 }
 
+// diskOf names the physical drive a mount's source device sits on — "/dev/nvme0n1p2" is on
+// "nvme0n1". It follows one layer of device-mapper or md (LUKS on a partition, say) but
+// gives up when a thing spans several drives, because then no single drive is the answer.
+func diskOf(source string) string {
+	if !strings.HasPrefix(source, "/dev/") {
+		return "" // a ZFS pool, a network share, a tmpfs: not a block device at all
+	}
+	d := parentDisk(strings.TrimPrefix(source, "/dev/"), 0)
+	// the sampler reports one drive per NVMe controller, so name it the same way
+	if m := nsRe.FindStringSubmatch(d); m != nil {
+		return m[1]
+	}
+	return d
+}
+
+func parentDisk(name string, depth int) string {
+	if name == "" || depth > 4 || strings.Contains(name, "/") {
+		return ""
+	}
+	base := "/sys/class/block/" + name
+	if _, err := os.Stat(base); err != nil {
+		return ""
+	}
+	// a partition: its parent directory in sysfs is the whole drive
+	if _, err := os.Stat(base + "/partition"); err == nil {
+		if target, err := os.Readlink(base); err == nil {
+			return filepath.Base(filepath.Dir(target))
+		}
+		return ""
+	}
+	// device-mapper or md: one drive underneath is an answer, several is not
+	if ents, err := os.ReadDir(base + "/slaves"); err == nil && len(ents) > 0 {
+		if len(ents) > 1 {
+			return ""
+		}
+		return parentDisk(ents[0].Name(), depth+1)
+	}
+	return name // a whole drive, used without partitioning
+}
+
 // ---- UPS ----
 
 func runOut(args ...string) (string, bool) {

internal/metrics/sensors.go

@@ -123,6 +123,9 @@ func flatten(s Sample, want func(id string) bool) []sensors.Sensor {
 		if x.Remote {
 			lbl = append(lbl, "remote", "1")
 		}
+		if x.Disk != "" {
+			lbl = append(lbl, "disk", x.Disk)
+		}
 		f.add(p+".size", x.Mount+" size", "bytes", "B", float64(x.Size), s.fsT, none, lbl...)
 		f.add(p+".used", x.Mount+" used", "bytes", "B", float64(x.Used), s.fsT, upTo(float64(x.Size)), lbl...)
 		f.add(p+".used_pct", x.Mount+" used", "usage", "%", ratio(x.Used, x.Size), s.fsT, pct, lbl...)

internal/space/walk.go

@@ -1,4 +1,6 @@
-// Package space measures what is using a disk: a parallel walk of the filesystems a disk
+// Package space measures what is using a disk — it is what the UI calls the **Usage** tab
+// (renamed from "Space" on 2026-09-24; the package and its /api/space endpoints kept the
+// old name rather than churn a working API): a parallel walk of the filesystems a disk
 // backs, kept as a tree of directories and the files worth seeing, the way WizTree does it.
 //
 // The walk never crosses a filesystem boundary by itself — /proc, /sys, another disk's

internal/storage/snapshots.go

@@ -0,0 +1,476 @@
+package storage
+
+// Scheduled snapshots with a retention period: take a snapshot of a dataset on a schedule,
+// and drop the ones that have aged out.
+//
+// **The naming is zfSnap's, deliberately.** grid has run zfSnap v1.11.1 from root cron since
+// before this repo existed and it works well, so rather than invent a second convention and
+// have two tools disagree about what a snapshot is, this writes and reads exactly what
+// zfSnap does: `<dataset>@2026-09-24_02.30.01--6w`, with the retention encoded in the name.
+// That has three consequences worth knowing. Ham Lab can show and manage the snapshots
+// zfSnap already took. The two can run side by side without fighting, because both only
+// ever destroy snapshots that have passed the expiry written into their own name. And
+// changing a policy's retention does not retroactively re-date old snapshots — each one
+// carries the rule it was taken under, which is the behaviour you want from a backup.
+//
+// Two rules make this safe to run unattended. Only a snapshot whose name parses as a stamp
+// and a TTL is ever a candidate — anything taken by hand or named anything else is invisible
+// to it. And every policy can be asked what it would delete, before it deletes anything.
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"sort"
+	"strconv"
+	"strings"
+	"time"
+
+	"hamlab/internal/cron"
+	"hamlab/internal/priv"
+)
+
+// stampLayout is zfSnap's timestamp: 2026-09-24_02.30.01.
+const stampLayout = "2006-01-02_15.04.05"
+
+// ttlSep separates the stamp from the retention in a snapshot name, as zfSnap writes it.
+const ttlSep = "--"
+
+// Policy is a standing instruction: snapshot this dataset on this schedule, keep the
+// snapshots this long.
+type Policy struct {
+	ID      string `json:"id"`
+	Name    string `json:"name"`
+	Dataset string `json:"dataset"`
+	// Recursive takes one atomic snapshot of the dataset and everything under it (zfs -r).
+	Recursive bool `json:"recursive"`
+	// Every is hourly | daily | weekly | monthly | custom. At is "HH:MM" (the minute alone
+	// is used for hourly). Weekday is 0=Sunday, Day is 1–28. With "custom", Cron is a plain
+	// cron expression and the rest is ignored — the escape hatch for anything the four
+	// common shapes cannot say.
+	Every   string `json:"every"`
+	At      string `json:"at"`
+	Weekday int    `json:"weekday"`
+	Day     int    `json:"day"`
+	Cron    string `json:"cron,omitempty"`
+	// KeepFor is how long a snapshot lives: "6w", "30d", "12h". It is written into the
+	// snapshot's own name, zfSnap-style, so each one carries the rule it was taken under.
+	// KeepMin is a floor — that many newest snapshots survive however old they are, so a
+	// dataset is never left with none.
+	KeepFor string `json:"keep_for"`
+	KeepMin int    `json:"keep_min"`
+	Prefix  string `json:"prefix"`
+	Enabled bool   `json:"enabled"`
+
+}
+
+func (p *Policy) prefix() string { return strings.TrimSpace(p.Prefix) }
+
+// Validate reports what is wrong with a policy, or nil.
+func (p *Policy) Validate() error {
+	if !datasetRe.MatchString(p.Dataset) {
+		return errors.New("dataset: a pool or pool/dataset name")
+	}
+	if pre := p.prefix(); pre != "" && (!snapRe.MatchString(pre) || strings.Contains(pre, ttlSep)) {
+		return errors.New("prefix: letters, digits, . _ : and - only, and no --")
+	}
+	switch p.Every {
+	case "hourly", "daily", "weekly", "monthly":
+	case "custom":
+		if strings.TrimSpace(p.Cron) == "" {
+			return errors.New("a custom schedule needs a cron expression")
+		}
+	default:
+		return errors.New("every: hourly, daily, weekly, monthly or custom")
+	}
+	if _, err := p.CronExpr(); err != nil {
+		return err
+	}
+	if _, _, err := p.clock(); err != nil {
+		return err
+	}
+	if p.Every == "weekly" && (p.Weekday < 0 || p.Weekday > 6) {
+		return errors.New("weekday: 0 (Sunday) to 6")
+	}
+	if p.Every == "monthly" && (p.Day < 1 || p.Day > 28) {
+		return errors.New("day: 1 to 28, so every month has one")
+	}
+	if p.KeepFor != "" {
+		if _, err := ParseAge(p.KeepFor); err != nil {
+			return err
+		}
+	}
+	if p.KeepMin < 0 {
+		return errors.New("keep at least: 0 or more")
+	}
+	if p.KeepFor == "" {
+		return errors.New("a retention period is required — it is written into each snapshot's name, and is what lets it be cleaned up later")
+	}
+	return nil
+}
+
+func (p *Policy) clock() (hour, min int, err error) {
+	at := strings.TrimSpace(p.At)
+	if at == "" {
+		return 0, 0, nil
+	}
+	h, m, ok := strings.Cut(at, ":")
+	if !ok {
+		return 0, 0, errors.New("at: HH:MM")
+	}
+	hour, err = strconv.Atoi(h)
+	if err == nil {
+		min, err = strconv.Atoi(m)
+	}
+	if err != nil || hour < 0 || hour > 23 || min < 0 || min > 59 {
+		return 0, 0, errors.New("at: HH:MM, 00:00 to 23:59")
+	}
+	return hour, min, nil
+}
+
+// ParseAge reads "6w", "30d", "12h", "90m" as a duration. Weeks and days are what a
+// retention period is actually written in, and time.ParseDuration knows neither.
+func ParseAge(s string) (time.Duration, error) {
+	s = strings.TrimSpace(strings.ToLower(s))
+	if s == "" {
+		return 0, errors.New("empty period")
+	}
+	unit := s[len(s)-1]
+	n, err := strconv.Atoi(s[:len(s)-1])
+	if err != nil || n <= 0 {
+		return 0, fmt.Errorf("%q: a number and a unit, like 6w, 30d or 12h", s)
+	}
+	switch unit {
+	case 'm':
+		return time.Duration(n) * time.Minute, nil
+	case 'h':
+		return time.Duration(n) * time.Hour, nil
+	case 'd':
+		return time.Duration(n) * 24 * time.Hour, nil
+	case 'w':
+		return time.Duration(n) * 7 * 24 * time.Hour, nil
+	}
+	return 0, fmt.Errorf("%q: the unit is m, h, d or w", s)
+}
+
+// CronExpr renders the schedule as a cron expression — which is what actually runs it. The
+// four common shapes are sugar over this; a custom policy states it outright. One expression
+// is the only description of when a schedule fires, so the page, the cron file and the next
+// run can never disagree.
+func (p *Policy) CronExpr() (string, error) {
+	if p.Every == "custom" {
+		expr := strings.Join(strings.Fields(p.Cron), " ")
+		if err := cron.ValidSchedule(expr); err != nil {
+			return "", fmt.Errorf("cron expression: %w", err)
+		}
+		if strings.HasPrefix(expr, "@reboot") {
+			return "", errors.New("@reboot is not a schedule for snapshots")
+		}
+		return expr, nil
+	}
+	hour, min, err := p.clock()
+	if err != nil {
+		return "", err
+	}
+	switch p.Every {
+	case "hourly":
+		return fmt.Sprintf("%d * * * *", min), nil
+	case "daily":
+		return fmt.Sprintf("%d %d * * *", min, hour), nil
+	case "weekly":
+		return fmt.Sprintf("%d %d * * %d", min, hour, p.Weekday), nil
+	case "monthly":
+		return fmt.Sprintf("%d %d %d * *", min, hour, p.Day), nil
+	}
+	return "", errors.New("unknown schedule")
+}
+
+// NextRun is when cron will next fire this schedule, read with the same evaluator the Jobs
+// tab uses on everyone else's lines.
+func (p *Policy) NextRun(now time.Time) time.Time {
+	expr, err := p.CronExpr()
+	if err != nil {
+		return time.Time{}
+	}
+	t, ok := cron.NextRun(expr, now)
+	if !ok {
+		return time.Time{}
+	}
+	return t
+}
+
+// SnapTime reads a snapshot's name in zfSnap's convention and returns when it was taken,
+// when it expires, and whether the name parsed at all. A name without a TTL never expires.
+// An optional prefix in front of the stamp is allowed, so a policy can label its own.
+func SnapTime(name string) (taken time.Time, expires time.Time, ok bool) {
+	_, snap, cut := strings.Cut(name, "@")
+	if !cut {
+		return
+	}
+	body, ttl, hasTTL := strings.Cut(snap, ttlSep)
+	// an optional "<prefix>-" before the stamp
+	stamp := body
+	if len(body) > len(stampLayout) {
+		stamp = body[len(body)-len(stampLayout):]
+	}
+	t, err := time.ParseInLocation(stampLayout, stamp, time.Local)
+	if err != nil {
+		return time.Time{}, time.Time{}, false
+	}
+	taken, ok = t, true
+	if hasTTL {
+		if d, err := ParseAge(ttl); err == nil {
+			expires = t.Add(d)
+		}
+	}
+	return
+}
+
+// Doomed is which of this dataset's snapshots have passed the expiry written into their own
+// names, oldest first — exactly what zfSnap -d would remove. This is what the UI shows
+// before anything is deleted, and what Run then acts on. KeepMin holds back that many of the
+// newest whatever their names say, so a dataset is never left with none.
+func (m *Manager) Doomed(ctx context.Context, p *Policy) ([]Snapshot, error) {
+	if err := p.Validate(); err != nil {
+		return nil, err
+	}
+	all, err := m.Snapshots(ctx, p.Dataset)
+	if err != nil {
+		return nil, err
+	}
+	var dated []Snapshot
+	expiry := map[string]time.Time{}
+	for _, s := range all {
+		taken, exp, ok := SnapTime(s.Name)
+		if !ok || exp.IsZero() {
+			continue // hand-made, or kept forever: not ours to remove
+		}
+		_ = taken
+		dated = append(dated, s)
+		expiry[s.Name] = exp
+	}
+	sort.Slice(dated, func(i, j int) bool { return dated[i].Created > dated[j].Created }) // newest first
+	now := time.Now()
+	var doomed []Snapshot
+	for i, s := range dated {
+		if i < p.KeepMin {
+			continue // the floor: these survive however old they are
+		}
+		if expiry[s.Name].Before(now) {
+			doomed = append(doomed, s)
+		}
+	}
+	sort.Slice(doomed, func(i, j int) bool { return doomed[i].Created < doomed[j].Created }) // oldest first
+	return doomed, nil
+}
+
+// DestroySnapshot removes one snapshot. recursive also removes the same-named snapshot on
+// every dataset beneath it, which is what a recursive policy created.
+func (m *Manager) DestroySnapshot(ctx context.Context, full string, recursive bool) error {
+	dataset, snap, ok := strings.Cut(full, "@")
+	if !ok || !datasetRe.MatchString(dataset) || !snapRe.MatchString(snap) {
+		return errors.New("bad snapshot name")
+	}
+	args := []string{"zfs", "destroy"}
+	if recursive {
+		args = append(args, "-r")
+	}
+	out, code, err := m.run(ctx, append(args, full)...)
+	if err != nil {
+		return err
+	}
+	if code != 0 {
+		return errors.New(strings.TrimSpace(string(out)))
+	}
+	return nil
+}
+
+// SnapshotRecursive takes dataset@name, optionally of everything beneath it too.
+func (m *Manager) SnapshotRecursive(ctx context.Context, dataset, name string, recursive bool) error {
+	if !datasetRe.MatchString(dataset) || !snapRe.MatchString(name) {
+		return errors.New("bad snapshot name")
+	}
+	args := []string{"zfs", "snapshot"}
+	if recursive {
+		args = append(args, "-r")
+	}
+	out, code, err := m.run(ctx, append(args, dataset+"@"+name)...)
+	if err != nil {
+		return err
+	}
+	if code != 0 {
+		return errors.New(strings.TrimSpace(string(out)))
+	}
+	return nil
+}
+
+// RunResult is what one pass of a policy did.
+type RunResult struct {
+	Taken   string   `json:"taken"`
+	Pruned  []string `json:"pruned"`
+	Skipped int      `json:"skipped"` // snapshots that are not this policy's, left alone
+}
+
+// Run takes the snapshot and prunes what has aged out. It is the same code the scheduler
+// calls, so "Run now" and the timer cannot drift apart.
+func (m *Manager) Run(ctx context.Context, p *Policy) (RunResult, error) {
+	var res RunResult
+	if err := p.Validate(); err != nil {
+		return res, err
+	}
+	name := time.Now().Format(stampLayout)
+	if pre := strings.TrimSpace(p.Prefix); pre != "" {
+		name = pre + "-" + name
+	}
+	if p.KeepFor != "" {
+		name += ttlSep + strings.TrimSpace(strings.ToLower(p.KeepFor)) // zfSnap's convention: the retention lives in the name
+	}
+	if err := m.SnapshotRecursive(ctx, p.Dataset, name, p.Recursive); err != nil {
+		return res, err
+	}
+	res.Taken = p.Dataset + "@" + name
+	doomed, err := m.Doomed(ctx, p)
+	if err != nil {
+		return res, err // the snapshot was taken; the prune is what failed
+	}
+	for _, s := range doomed {
+		if err := m.DestroySnapshot(ctx, s.Name, p.Recursive); err != nil {
+			return res, fmt.Errorf("destroying %s: %w", s.Name, err)
+		}
+		res.Pruned = append(res.Pruned, s.Name)
+	}
+	return res, nil
+}
+
+// ---- what happened when it ran ----
+//
+// The policies are a definition the daemon owns; what a run did is written by whatever ran
+// it, which is a cron job in its own process. Keeping the two apart means the CLI never has
+// to write state.json behind the daemon's back — the last writer would win and one of them
+// would lose its work.
+
+// RunRecord is the outcome of one run of one schedule.
+type RunRecord struct {
+	At     time.Time `json:"at"`
+	Taken  string    `json:"taken,omitempty"`
+	Pruned []string  `json:"pruned,omitempty"`
+	Error  string    `json:"error,omitempty"`
+}
+
+func runPath(dir, id string) string { return filepath.Join(dir, "zfs-runs", id+".json") }
+
+// WriteRun records what a run did, beside the config.
+func WriteRun(dir, id string, r RunRecord) error {
+	if err := os.MkdirAll(filepath.Join(dir, "zfs-runs"), 0o700); err != nil {
+		return err
+	}
+	b, err := json.Marshal(r)
+	if err != nil {
+		return err
+	}
+	p := runPath(dir, id)
+	tmp := p + ".tmp"
+	if err := os.WriteFile(tmp, b, 0o600); err != nil {
+		return err
+	}
+	return os.Rename(tmp, p)
+}
+
+// ReadRun is the last thing a schedule did, or the zero value if it has never run.
+func ReadRun(dir, id string) RunRecord {
+	var r RunRecord
+	b, err := os.ReadFile(runPath(dir, id))
+	if err == nil {
+		_ = json.Unmarshal(b, &r)
+	}
+	return r
+}
+
+// ForgetRun drops a schedule's record, for when the schedule itself goes.
+func ForgetRun(dir, id string) { _ = os.Remove(runPath(dir, id)) }
+
+// ---- the cron file ----
+
+// CronFile is where Ham Lab writes the snapshot schedules. One file it owns entirely and
+// rewrites whole, so there is never a half-updated set of lines, and so a schedule removed
+// in the UI is removed from the machine.
+const CronFile = "/etc/cron.d/hamlab-zfs"
+
+// CronFileText renders every schedule as cron lines. user is who they run as — the account
+// Ham Lab itself runs as, which reaches zfs through sudo exactly as the daemon does, so the
+// files it writes stay owned by that account rather than by root.
+func CronFileText(policies []Policy, user, exe, config string) (string, error) {
+	var b strings.Builder
+	b.WriteString("# Ham Lab snapshot schedules — written by the ZFS tab, rewritten whole on every change.\n")
+	b.WriteString("# Edit them there, not here: anything added by hand is lost the next time a schedule changes.\n")
+	b.WriteString("SHELL=/bin/sh\n")
+	b.WriteString("PATH=/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin\n\n")
+	for i := range policies {
+		p := &policies[i]
+		if !p.Enabled {
+			b.WriteString("# " + p.Name + " — paused\n")
+			continue
+		}
+		expr, err := p.CronExpr()
+		if err != nil {
+			return "", fmt.Errorf("%s: %w", p.Name, err)
+		}
+		name := strings.ReplaceAll(p.Name, "\n", " ")
+		fmt.Fprintf(&b, "# %s — %s, keep %s\n", name, p.Dataset, p.KeepFor)
+		fmt.Fprintf(&b, "%s %s %s -config %s zfs-snapshot %s\n\n", expr, user, exe, config, p.ID)
+	}
+	return b.String(), nil
+}
+
+// WriteCronFile puts the schedules on the machine, or removes the file when there are none.
+func (m *Manager) WriteCronFile(ctx context.Context, policies []Policy, user, exe, config string) error {
+	live := false
+	for i := range policies {
+		if policies[i].Enabled {
+			live = true
+		}
+	}
+	if !live {
+		out, code, err := m.run(ctx, "rm", "-f", CronFile)
+		if err != nil {
+			return err
+		}
+		if code != 0 {
+			return errors.New(strings.TrimSpace(string(out)))
+		}
+		return nil
+	}
+	text, err := CronFileText(policies, user, exe, config)
+	if err != nil {
+		return err
+	}
+	return m.writeRoot(ctx, CronFile, text)
+}
+
+// writeRoot puts text at a root-owned path, through the same privilege escalation every
+// other write here uses.
+func (m *Manager) writeRoot(ctx context.Context, path, text string) error {
+	args := []string{"tee", path}
+	if m.sudo {
+		args = priv.Wrap(args...)
+	}
+	ctx, cancel := context.WithTimeout(ctx, 30*time.Second)
+	defer cancel()
+	cmd := exec.CommandContext(ctx, args[0], args[1:]...)
+	cmd.Stdin = strings.NewReader(text)
+	var errb bytes.Buffer
+	cmd.Stdout, cmd.Stderr = nil, &errb
+	if err := cmd.Run(); err != nil {
+		msg := strings.TrimSpace(errb.String())
+		if msg == "" {
+			msg = err.Error()
+		}
+		return errors.New(msg)
+	}
+	return nil
+}

internal/storage/snapshots_test.go

@@ -0,0 +1,135 @@
+package storage
+
+import (
+	"strings"
+	"testing"
+	"time"
+
+	"hamlab/internal/cron"
+)
+
+func TestSnapTimeReadsZfSnapNames(t *testing.T) {
+	// the real thing off grid
+	taken, exp, ok := SnapTime("tank@2026-09-24_02.30.01--6w")
+	if !ok {
+		t.Fatal("zfSnap's own name must parse")
+	}
+	if taken.Format("2006-01-02 15:04:05") != "2026-09-24 02:30:01" {
+		t.Fatalf("taken = %v", taken)
+	}
+	if got := exp.Sub(taken); got != 6*7*24*time.Hour {
+		t.Fatalf("ttl = %v, want 6w", got)
+	}
+	// a labelled one from a policy with a prefix
+	if _, e, ok := SnapTime("tank/apps@nightly-2026-09-24_02.30.01--30d"); !ok || e.IsZero() {
+		t.Fatal("a prefixed name must parse and expire")
+	}
+	// no TTL: never expires
+	if _, e, ok := SnapTime("tank@2026-09-24_02.30.01"); !ok || !e.IsZero() {
+		t.Fatal("a name without a TTL must parse and never expire")
+	}
+	// things that are not ours
+	for _, n := range []string{"tank@before-upgrade", "tank@manual", "tank", "tank@2026-13-45_99.99.99--6w"} {
+		if _, _, ok := SnapTime(n); ok {
+			t.Errorf("%q must not parse as a scheduled snapshot", n)
+		}
+	}
+}
+
+func TestCronExprAndNextRun(t *testing.T) {
+	now := time.Date(2026, 9, 24, 9, 15, 0, 0, time.Local)
+	for _, c := range []struct {
+		p    Policy
+		want string
+	}{
+		{Policy{Every: "daily", At: "02:30"}, "30 2 * * *"},
+		{Policy{Every: "hourly", At: "00:30"}, "30 * * * *"},
+		{Policy{Every: "weekly", At: "04:00", Weekday: 0}, "0 4 * * 0"},
+		{Policy{Every: "monthly", At: "05:15", Day: 12}, "15 5 12 * *"},
+		{Policy{Every: "custom", Cron: "*/15  9-17 * * 1-5"}, "*/15 9-17 * * 1-5"},
+	} {
+		got, err := c.p.CronExpr()
+		if err != nil || got != c.want {
+			t.Errorf("CronExpr(%+v) = %q, %v; want %q", c.p.Every, got, err, c.want)
+		}
+	}
+	// the next run is read with the same evaluator the Jobs tab uses
+	daily := Policy{Every: "daily", At: "02:30"}
+	if n := daily.NextRun(now); n.Hour() != 2 || n.Minute() != 30 || n.Day() != 25 {
+		t.Fatalf("NextRun = %v, want 25th 02:30", n)
+	}
+	// a custom expression that does not parse is refused, not silently ignored
+	if _, err := (&Policy{Every: "custom", Cron: "not a schedule"}).CronExpr(); err == nil {
+		t.Fatal("a bad cron expression must be refused")
+	}
+	if _, err := (&Policy{Every: "custom", Cron: "@reboot"}).CronExpr(); err == nil {
+		t.Fatal("@reboot is not a snapshot schedule")
+	}
+}
+
+func TestCronFileText(t *testing.T) {
+	ps := []Policy{
+		{ID: "abc", Name: "tank nightly", Dataset: "tank", Every: "daily", At: "02:30", KeepFor: "6w", Enabled: true},
+		{ID: "def", Name: "paused one", Dataset: "tank/apps", Every: "hourly", At: "00:05", KeepFor: "2d"},
+	}
+	text, err := CronFileText(ps, "core", "/usr/local/bin/hamlab", "/home/core/apps/hamlab/config.json")
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !strings.Contains(text, "30 2 * * * core /usr/local/bin/hamlab -config /home/core/apps/hamlab/config.json zfs-snapshot abc") {
+		t.Fatalf("the enabled schedule is missing or malformed:\n%s", text)
+	}
+	if strings.Contains(text, "zfs-snapshot def") {
+		t.Fatalf("a paused schedule must not be scheduled:\n%s", text)
+	}
+	// every line cron will read must parse as a schedule
+	for _, line := range strings.Split(text, "\n") {
+		line = strings.TrimSpace(line)
+		if line == "" || strings.HasPrefix(line, "#") || strings.Contains(line, "=") {
+			continue
+		}
+		f := strings.Fields(line)
+		if err := cron.ValidSchedule(strings.Join(f[:5], " ")); err != nil {
+			t.Errorf("line %q: %v", line, err)
+		}
+	}
+}
+
+func TestValidateRefusesNonsense(t *testing.T) {
+	ok := Policy{Dataset: "tank/apps", Every: "daily", At: "02:30", KeepFor: "6w"}
+	if err := ok.Validate(); err != nil {
+		t.Fatalf("a sane policy: %v", err)
+	}
+	bad := []Policy{
+		{Dataset: "", Every: "daily", At: "02:30", KeepFor: "6w"},
+		{Dataset: "tank", Every: "yearly", At: "02:30", KeepFor: "6w"},
+		{Dataset: "tank", Every: "custom", KeepFor: "6w"},
+		{Dataset: "tank", Every: "custom", Cron: "not a schedule", KeepFor: "6w"},
+		{Dataset: "tank", Every: "daily", At: "25:00", KeepFor: "6w"},
+		{Dataset: "tank", Every: "daily", At: "02:30"},                          // no retention
+		{Dataset: "tank", Every: "daily", At: "02:30", KeepFor: "6 weeks"},      // unparseable
+		{Dataset: "tank", Every: "monthly", At: "02:30", Day: 31, KeepFor: "6w"}, // not every month has one
+		{Dataset: "tank", Every: "daily", At: "02:30", KeepFor: "6w", Prefix: "a--b"},
+	}
+	for i, p := range bad {
+		if err := p.Validate(); err == nil {
+			t.Errorf("bad policy %d was accepted: %+v", i, p)
+		}
+	}
+}
+
+func TestParseAge(t *testing.T) {
+	for s, want := range map[string]time.Duration{
+		"6w": 6 * 7 * 24 * time.Hour, "30d": 30 * 24 * time.Hour,
+		"12h": 12 * time.Hour, "90m": 90 * time.Minute,
+	} {
+		if got, err := ParseAge(s); err != nil || got != want {
+			t.Errorf("ParseAge(%q) = %v, %v; want %v", s, got, err, want)
+		}
+	}
+	for _, s := range []string{"", "6", "w", "-1d", "0w", "6y", "six weeks"} {
+		if _, err := ParseAge(s); err == nil {
+			t.Errorf("ParseAge(%q) should fail", s)
+		}
+	}
+}

internal/store/store.go

@@ -51,7 +51,8 @@ type State struct {
 	// UI is the pre-users machine-wide look; on first boot with users it moves to the first
 	// account. Kept so an old state file still reads.
 	UI         json.RawMessage `json:"ui,omitempty"`
-	Alerts     json.RawMessage `json:"alerts,omitempty"`      // rules, channels and recent events (internal/api/alerts.go)
+	Alerts     json.RawMessage `json:"alerts,omitempty"`
+	ZFS        json.RawMessage `json:"zfs,omitempty"`  // snapshot schedules (internal/api/zfs.go)      // rules, channels and recent events (internal/api/alerts.go)
 	SamplerMs  int             `json:"sampler_ms,omitempty"`  // how often this machine samples; the hub keeps every machine on the console's refresh rate
 	HistorySec int             `json:"history_sec,omitempty"` // how far back the graphs reach; the hub keeps every machine on the same one
 }

web/src/Disks.tsx

@@ -1,8 +1,8 @@
 import { useCallback, useEffect, useMemo, useState } from 'react'
-import { api, SmartInfo, StorageDevice, ZFSInfo, ZFSSnapshot } from './api'
+import { api, SmartInfo, StorageDevice } from './api'
 import { fmtBytes, fmtRate } from './fmt'
 import TaskOutput from './TaskOutput'
-import { Btn, Loading, Spinner, askText, toast } from './ui'
+import { askText, Btn, Loading, Meter, Spinner, toast } from './ui'
 import { useRefreshMs } from './prefs'
 import { Sensor, useSensorStream } from './sensors'
 
@@ -11,10 +11,6 @@ const KIND: Record<string, string> = { nvme: 'NVMe', ssd: 'SSD', hdd: 'HDD', usb
 const hours = (h: number) => (h >= 24 * 365 ? `${(h / 8760).toFixed(1)} y` : h >= 48 ? `${Math.round(h / 24)} d` : `${h} h`)
 const ago = (ms: number) => { const s = (Date.now() - ms) / 1000; return s < 3600 ? `${Math.round(s / 60)}m ago` : s < 86400 ? `${Math.round(s / 3600)}h ago` : `${Math.round(s / 86400)}d ago` }
 
-function Meter({ frac, cls }: { frac: number; cls?: string }) {
-  const c = cls ?? (frac >= 0.95 ? 'bad' : frac >= 0.85 ? 'warn' : '')
-  return <div className="meter"><div className={'fill ' + c} style={{ width: `${Math.min(100, Math.max(0, frac * 100))}%` }} /></div>
-}
 
 // ---- SMART panel ----
 function SmartPanel({ dev, rotational }: { dev: string; rotational: boolean }) {
@@ -109,82 +105,9 @@ function DeviceCard({ d, live, forceOpen }: { d: StorageDevice; live?: { read_bp
   )
 }
 
-// ---- ZFS ----
-function PoolCard({ p, onChanged }: { p: ZFSInfo['pools'][number]; onChanged: () => void }) {
-  const scrubbing = /scrub in progress/i.test(p.scan ?? '')
-  const paused = /scrub paused/i.test(p.scan ?? '')
-  const [cfg, setCfg] = useState(false)
-  const act = (a: 'start' | 'pause' | 'stop') => async () => {
-    try { await api.zfsScrub(p.name, a); toast(`scrub ${a}: ok`); onChanged() } catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
-  }
-  return (
-    <section className="card">
-      <div className="sub-head">
-        <span className={'dot ' + (p.health === 'ONLINE' ? 'ok' : p.health === 'DEGRADED' ? 'warn' : 'bad')} />
-        <span className="dname">{p.name}</span>
-        <span className="badge">{p.health}</span>
-        <span className="spacer" />
-        <span className="num">{fmtBytes(p.alloc)} / {fmtBytes(p.size)} · {p.cap}%</span>
-      </div>
-      <Meter frac={p.cap / 100} cls={p.cap >= 90 ? 'bad' : p.cap >= 80 ? 'warn' : ''} />
-      <div className="props" style={{ marginTop: 8 }}>
-        <div><span className="k">Free</span><span>{fmtBytes(p.free)} · fragmentation {p.frag}% · dedup {p.dedup}</span></div>
-        {p.scan && <div><span className="k">Scan</span><span style={{ whiteSpace: 'normal' }}>{p.scan}</span></div>}
-        {p.errors && <div><span className="k">Errors</span><span className={p.errors.startsWith('No known') ? '' : 'warn-text'}>{p.errors}</span></div>}
-      </div>
-      <div className="svc-actions">
-        {!scrubbing && <Btn className="btn" confirm={`Start a scrub of ${p.name}? It reads the whole pool and can take many hours.`} onClick={act('start')}>{paused ? 'Resume scrub' : 'Scrub'}</Btn>}
-        {scrubbing && <Btn className="btn" onClick={act('pause')}>Pause scrub</Btn>}
-        {(scrubbing || paused) && <Btn className="btn" confirm="Stop the scrub?" onClick={act('stop')}>Stop scrub</Btn>}
-        <button className="btn" onClick={() => setCfg(!cfg)}>{cfg ? 'Hide vdevs' : 'Vdevs'}</button>
-      </div>
-      {cfg && <pre className="unit-file">{p.config}</pre>}
-    </section>
-  )
-}
-
-function DatasetRow({ d, onChanged }: { d: ZFSInfo['datasets'][number]; onChanged: () => void }) {
-  const [snaps, setSnaps] = useState<ZFSSnapshot[] | null>(null)
-  const [open, setOpen] = useState(false)
-  async function toggle() {
-    setOpen(!open)
-    if (!open && !snaps) { try { setSnaps(await api.zfsSnapshots(d.name)) } catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err'); setSnaps([]) } }
-  }
-  async function snapshot() {
-    const def = 'hamlab-' + new Date().toISOString().slice(0, 16).replace(/[-:T]/g, '').replace(/(\d{8})(\d{4})/, '$1-$2')
-    const name = await askText(`Snapshot name for ${d.name}@`, def)
-    if (!name) return
-    try { await api.zfsSnapshot(d.name, name.trim()); toast(`${d.name}@${name.trim()} created`); setSnaps(null); onChanged() }
-    catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
-  }
-  const depth = d.name.split('/').length - 1
-  return (
-    <>
-      <div className="ds-row" style={{ paddingLeft: 12 + depth * 16 }}>
-        <span className="ds-name" onClick={toggle}><span className="chev">{open ? '▾' : '▸'}</span>{d.name.split('/').pop()}{d.type === 'volume' && <span className="badge">zvol</span>}</span>
-        <span className="num">{fmtBytes(d.used)}</span>
-        <span className="num muted">{fmtBytes(d.refer)}</span>
-        <span className="num muted">{d.used_by_snapshots ? fmtBytes(d.used_by_snapshots) : '—'}</span>
-        <span className="num muted">{d.snapshots}</span>
-        <span className="num muted">{d.compressratio}</span>
-        <span className="ds-mount muted">{d.mount}</span>
-      </div>
-      {open && (
-        <div className="ds-snaps">
-          <div className="svc-actions"><Btn className="btn" onClick={snapshot}>Snapshot now</Btn><span className="muted">avail {fmtBytes(d.avail)}</span></div>
-          {snaps === null ? <div className="muted"><Spinner /> listing…</div> : snaps.length === 0 ? <div className="muted">No snapshots.</div> : (
-            <div className="snap-list">{snaps.slice().reverse().map((s) => <div key={s.name} className="snap"><span className="mono">{s.name.split('@')[1]}</span><span className="num muted">{new Date(s.created).toLocaleString()} · {fmtBytes(s.used)}</span></div>)}</div>
-          )}
-        </div>
-      )}
-    </>
-  )
-}
-
 // ---- page ----
 export default function Disks() {
   const [devs, setDevs] = useState<StorageDevice[] | null>(null)
-  const [zfs, setZfs] = useState<ZFSInfo | null>(null)
   const [hasSmart, setHasSmart] = useState(true)
   const [err, setErr] = useState('')
   const [task, setTask] = useState<string | null>(null)
@@ -194,7 +117,7 @@ export default function Disks() {
   const load = useCallback(async () => {
     try {
       const r = await api.storage()
-      setDevs(r.devices); setZfs(r.zfs); setHasSmart(r.smartctl); setErr('')
+      setDevs(r.devices); setHasSmart(r.smartctl); setErr('')
     } catch (e) { setErr(e instanceof Error ? e.message : 'failed') }
   }, [])
   useEffect(() => { load() }, [load])
@@ -237,18 +160,6 @@ export default function Disks() {
         {devs.map((d) => <DeviceCard key={d.name} d={d} live={liveFor(d)} forceOpen={allSmart ? true : null} />)}
       </div>
 
-      <div className="dash-bar"><h2 className="section">ZFS</h2></div>
-      {zfs && !zfs.available && <div className="notice">{zfs.message || 'No ZFS pools on this host.'}</div>}
-      {zfs?.available && zfs.pools.length === 0 && <div className="notice">ZFS is present but no pools are imported.</div>}
-      {zfs?.available && zfs.pools.length > 0 && (
-        <>
-          <div className="cards">{zfs.pools.map((p) => <PoolCard key={p.name} p={p} onChanged={load} />)}</div>
-          <section className="card datasets">
-            <div className="ds-row head"><span>Dataset</span><span className="num">Used</span><span className="num">Refer</span><span className="num">In snaps</span><span className="num">Snaps</span><span className="num">Ratio</span><span>Mount</span></div>
-            {zfs.datasets.map((d) => <DatasetRow key={d.name} d={d} onChanged={load} />)}
-          </section>
-        </>
-      )}
     </div>
   )
 }

web/src/Jobs.tsx

@@ -175,11 +175,11 @@ function TimerDetail({ t, onChanged, onBack }: { t: CronTimer; onChanged: () =>
 }
 
 // ---- page ----
-export default function Cron() {
+export default function Jobs() {
   const [ov, setOv] = useState<CronOverview | null>(null)
   const [err, setErr] = useState('')
   const [q, setQ] = useState('')
-  const [hashSel, setHashSel] = useHashSel('cron') // "job:<id>" or "timer:<unit>", in the URL
+  const [hashSel, setHashSel] = useHashSel('jobs') // "job:<id>" or "timer:<unit>", in the URL
   const sel = /^(job|timer):./.test(hashSel) ? { kind: hashSel.split(':')[0] as 'job' | 'timer', id: hashSel.slice(hashSel.indexOf(':') + 1) } : null
   const setSel = (v: { kind: 'job' | 'timer'; id: string } | null) => setHashSel(v ? `${v.kind}:${v.id}` : '')
   const [editing, setEditing] = useState<string | null>(null)
@@ -206,7 +206,7 @@ export default function Cron() {
   else pane = <div className="svc-empty muted">Select a job or timer.</div>
 
   return (
-    <div className={'cron' + (sel || editing ? ' has-sel' : '')}>
+    <div className={'cron jobs' + (sel || editing ? ' has-sel' : '')}>
       <div className="cron-list">
         <div className="svc-tools">
           <div className="tools-row"><input placeholder="filter jobs and timers" value={q} onChange={(e) => setQ(e.target.value)} /><Btn className="btn" title="reload" onClick={load}>↻</Btn></div>
@@ -218,7 +218,7 @@ export default function Cron() {
         {ov && !ov.logging && <div className="notice" style={{ margin: 8 }}>cron isn’t logging to the journal on this host, so “last run” is unavailable for cron jobs.</div>}
         <div className="cron-rows">
           {!ov && !err && <Loading label="Reading crontabs…" />}
-          {ov && <h3>Cron jobs · {jobs.length}</h3>}
+          {ov && <h3>Scheduled jobs · {jobs.length}</h3>}
           {jobs.map((j) => (
             <div key={j.id} className={'cron-row' + (sel?.kind === 'job' && sel.id === j.id ? ' active' : '') + (j.enabled ? '' : ' off')} onClick={() => { setEditing(null); setSel({ kind: 'job', id: j.id }) }}>
               <span className={'sdot ' + (j.enabled ? 'ok' : '')} />

web/src/KVM.tsx

@@ -0,0 +1,153 @@
+// IP-KVMs: a device's own web page, shown inside Ham Lab.
+//
+// Paste the address, and the device appears. Nothing here knows anything about any
+// particular KVM — it is an iframe pointed at Ham Lab's own proxy, so whatever web UI the
+// device serves is what you get.
+//
+// On the keyboard: a page cannot capture keys and forward them into an iframe, and does not
+// need to. An iframe with focus receives them directly. What the browser itself swallows —
+// Escape, Tab, Ctrl-W, the function keys — is what `navigator.keyboard.lock()` releases,
+// and that needs fullscreen and only exists in Chromium. Everywhere else you get the plain
+// version, which is most keys, and the device's own on-screen buttons for the rest.
+import { FormEvent, useCallback, useEffect, useRef, useState } from 'react'
+import { KVM as Device, MachineStatus, api } from './api'
+import { Btn, Loading, askYes, toast, useEscape } from './ui'
+
+// lock() releases the keys the browser would otherwise keep for itself. Chromium only; a
+// browser without it simply does not get the extra keys.
+type KeyboardLock = { lock?: (keys?: string[]) => Promise<void>; unlock?: () => void }
+const keyboardAPI = (): KeyboardLock | undefined => (navigator as unknown as { keyboard?: KeyboardLock }).keyboard
+
+export default function KVM({ machines }: { machines: MachineStatus[] }) {
+  const [list, setList] = useState<Device[] | null>(null)
+  const [sel, setSel] = useState<string>('')
+  const [adding, setAdding] = useState(false)
+  const [err, setErr] = useState('')
+  const [full, setFull] = useState(false)
+  const box = useRef<HTMLDivElement>(null)
+  const frame = useRef<HTMLIFrameElement>(null)
+
+  const load = useCallback(async () => {
+    try { const l = await api.kvms(); setList(l); setSel((s) => (l.some((k) => k.id === s) ? s : l[0]?.id ?? '')) }
+    catch (e) { setErr(e instanceof Error ? e.message : 'failed') }
+  }, [])
+  useEffect(() => { load() }, [load])
+
+  // Fullscreen plus a keyboard lock is what makes Escape and Tab reach the machine rather
+  // than the browser. Both are undone together when fullscreen ends, however it ends.
+  useEffect(() => {
+    const onChange = () => {
+      const on = !!document.fullscreenElement
+      setFull(on)
+      if (!on) keyboardAPI()?.unlock?.()
+    }
+    document.addEventListener('fullscreenchange', onChange)
+    return () => { document.removeEventListener('fullscreenchange', onChange); keyboardAPI()?.unlock?.() }
+  }, [])
+
+  const goFullscreen = async () => {
+    try {
+      await box.current?.requestFullscreen()
+      await keyboardAPI()?.lock?.()
+      frame.current?.focus()
+    } catch (e) { toast(e instanceof Error ? e.message : 'fullscreen refused', 'err') }
+  }
+
+  const remove = async (k: Device) => {
+    if (!await askYes(`Remove ${k.name}? The device is not touched — this only forgets its address.`, 'Remove')) return
+    try { await api.removeKVM(k.id); toast(`${k.name} removed`); await load() } catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
+  }
+
+  if (err && !list) return <div className="dash"><div className="error">{err}</div></div>
+  if (!list) return <div className="dash"><Loading label="Reading the list…" /></div>
+
+  const cur = list.find((k) => k.id === sel)
+  const locks = !!keyboardAPI()?.lock
+
+  return (
+    <div className="kvm">
+      <div className="kvm-bar">
+        {list.length > 0 && (
+          <span className="chips">
+            {list.map((k) => (
+              <button key={k.id} className={'chip' + (k.id === sel ? ' on' : '')} onClick={() => setSel(k.id)}
+                title={`${k.url}${k.machine ? ` — attached to ${k.machine}` : ''}`}>{k.name}</button>
+            ))}
+          </span>
+        )}
+        <span className="spacer" />
+        {cur && <>
+          <span className="muted small mono kvm-url">{cur.url}</span>
+          <Btn className="btn" onClick={goFullscreen} title={locks ? 'fullscreen, and the browser stops keeping Escape, Tab and Ctrl-W for itself' : 'fullscreen — this browser will still keep some keys for itself'}>Fullscreen</Btn>
+          <a className="btn" href={cur.url} target="_blank" rel="noreferrer" title="the device directly, outside Ham Lab">Open ↗</a>
+          <Btn className="btn danger" onClick={() => remove(cur)}>Remove</Btn>
+        </>}
+        <Btn className="btn on" onClick={() => setAdding(true)}>+ Add</Btn>
+      </div>
+
+      {cur ? (
+        <div className={'kvm-frame' + (full ? ' full' : '')} ref={box}>
+          <iframe ref={frame} key={cur.id} src={`/kvm/${cur.id}/`} title={cur.name}
+            allow="fullscreen; clipboard-read; clipboard-write" />
+        </div>
+      ) : (
+        <div className="svc-empty muted">
+          {list.length === 0
+            ? 'No KVM yet. Add one with its address and its web page appears here.'
+            : 'Pick a KVM.'}
+        </div>
+      )}
+      {cur && !full && (
+        <div className="kvm-hint muted small">
+          Click the picture, then type — the keys go to the machine.
+          {locks ? ' Fullscreen also takes Escape, Tab and Ctrl-W back from the browser.' : ' This browser keeps Escape, Tab and Ctrl-W for itself; use the device’s own on-screen keys for those, or a Chromium browser.'}
+        </div>
+      )}
+      {adding && <AddDialog machines={machines} onClose={() => setAdding(false)} onAdded={(k) => { setAdding(false); setSel(k.id); load() }} />}
+    </div>
+  )
+}
+
+function AddDialog({ machines, onClose, onAdded }: { machines: MachineStatus[]; onClose: () => void; onAdded: (k: Device) => void }) {
+  const [url, setUrl] = useState('')
+  const [name, setName] = useState('')
+  const [machine, setMachine] = useState('')
+  const [err, setErr] = useState('')
+  const [busy, setBusy] = useState(false)
+  useEscape(onClose)
+  const save = async (e: FormEvent) => {
+    e.preventDefault(); setErr(''); setBusy(true)
+    try { onAdded(await api.addKVM(name.trim(), url.trim(), machine)) }
+    catch (e) { setErr(e instanceof Error ? e.message : 'failed'); setBusy(false) }
+  }
+  return (
+    <div className="modal-back" onClick={onClose}>
+      <form className="modal" onClick={(e) => e.stopPropagation()} onSubmit={save}>
+        <h3>Add a KVM</h3>
+        <label className="muted">Address — whatever you would type into a browser
+          <input autoFocus className="mono" value={url} onChange={(e) => setUrl(e.target.value)} placeholder="10.0.0.217" spellCheck={false} autoCapitalize="off" required />
+        </label>
+        <label className="muted">Name — optional; the address is used otherwise
+          <input value={name} onChange={(e) => setName(e.target.value)} placeholder="mac mini KVM" maxLength={64} />
+        </label>
+        <label className="muted">Attached to — optional, and only a label
+          <select value={machine} onChange={(e) => setMachine(e.target.value)}>
+            <option value="">—</option>
+            {machines.map((m) => <option key={m.name} value={m.name}>{m.host || m.name}</option>)}
+          </select>
+        </label>
+        <div className="muted small">
+          Ham Lab serves the device from its own address, so a KVM that only speaks plain HTTP
+          still works here, and it is behind this login rather than needing a second one. The
+          device itself is not changed and is never reachable from outside.
+        </div>
+        {err && <div className="error">{err}</div>}
+        <div className="svc-actions">
+          <button type="submit" className="btn on" disabled={busy || !url.trim()}>Add</button>
+          <span className="spacer" />
+          <button type="button" className="btn" onClick={onClose}>Cancel</button>
+        </div>
+      </form>
+    </div>
+  )
+}

web/src/Shell.tsx

@@ -8,14 +8,16 @@ import Gallery from './Gallery'
 import Docker from './Docker'
 import Alerts from './Alerts'
 import Services from './Services'
-import Cron from './Cron'
+import Jobs from './Jobs'
 import Updates from './Updates'
 import Disks from './Disks'
+import ZFS from './ZFS'
+import KVM from './KVM'
 import Files from './Files'
 import Top from './Top'
 import Machines from './Machines'
 import Settings from './Settings'
-import Space from './Space'
+import Usage from './Usage'
 import Users from './Users'
 import MachineSwitcher from './MachineSwitcher'
 import { MachineStatus, setMachine } from './api'
@@ -54,28 +56,35 @@ const GROUPS: Group[] = [
   { id: 'terminal', label: 'Terminal', icon: '>_' },
   { id: 'services', label: 'Services', icon: '◈' },
   { id: 'docker', label: 'Docker', icon: DockerIcon },
-  { id: 'cron', label: 'Cron', icon: ClockIcon },
+  { id: 'jobs', label: 'Jobs', icon: ClockIcon },
   { id: 'updates', label: 'Updates', icon: '⇪' },
   { id: 'disks', label: 'Disks', icon: '▤' },
-  { id: 'space', label: 'Space', icon: '▧' },
+  { id: 'zfs', label: 'ZFS', icon: '⛁' },
+  { id: 'usage', label: 'Usage', icon: '▧' },
   { id: 'files', label: 'Files', icon: FolderIcon },
   { id: 'alerts', label: 'Alerts', icon: BellIcon },
   { id: 'machines', label: 'Machines', icon: '⌂' },
+  { id: 'kvm', label: 'KVM', icon: '⌨' },
   { id: 'users', label: 'Users', icon: '⚇' },
   { id: 'widgets', label: 'Widgets', icon: '▦' },
   { id: 'settings', label: 'Settings', icon: '⚙' },
 ]
 
-type View = 'terminal' | 'widgets' | 'docker' | 'alerts' | 'services' | 'cron' | 'updates' | 'disks' | 'space' | 'files' | 'procs' | 'machines' | 'users' | 'settings'
+type View = 'terminal' | 'widgets' | 'docker' | 'alerts' | 'services' | 'jobs' | 'updates' | 'disks' | 'zfs' | 'usage' | 'files' | 'procs' | 'machines' | 'kvm' | 'users' | 'settings'
-const VIEWS: View[] = ['terminal', 'widgets', 'docker', 'alerts', 'services', 'cron', 'updates', 'disks', 'space', 'files', 'procs', 'machines', 'users', 'settings']
+const VIEWS: View[] = ['terminal', 'widgets', 'docker', 'alerts', 'services', 'jobs', 'updates', 'disks', 'zfs', 'usage', 'files', 'procs', 'machines', 'kvm', 'users', 'settings']
-const HUB_ONLY: View[] = ['machines', 'users', 'settings', 'alerts']
+const HUB_ONLY: View[] = ['machines', 'kvm', 'users', 'settings', 'alerts']
 
 // URL hash is the source of truth for where you are: #/<machine|hub>/<view>. Back and
 // forward work, reload restores the tab, and links can point at a page on a machine.
+// Pages that were renamed, so a link or a bookmark made before it still lands where it meant.
+const RENAMED_VIEWS: Record<string, View> = { space: 'usage', cron: 'jobs', top: 'procs', dashboard: 'procs' }
+
 function parseHash(): { machine: string | null; view: View } | null {
   const m = location.hash.match(/^#\/([a-z0-9-]+)\/([a-z]+)(?:\/.*)?$/) // anything after the page is that page's own selection
-  if (!m || !VIEWS.includes(m[2] as View)) return null
+  if (!m) return null
-  return { machine: m[1] === 'hub' ? null : m[1], view: m[2] as View }
+  const view = VIEWS.includes(m[2] as View) ? (m[2] as View) : RENAMED_VIEWS[m[2]]
+  if (!view) return null
+  return { machine: m[1] === 'hub' ? null : m[1], view }
 }
 
 function fmtUptime(s: number): string {
@@ -262,16 +271,18 @@ export default function Shell({ me: hubMe, onLoggedOut }: { me: Me; onLoggedOut:
         <main className="main" key={machine ?? 'hub'}>
           {view === 'settings' && <Settings me={hubMe} />}
           {view === 'users' && <Users me={hubMe} />}
-          {view === 'space' && <Space key={machine ?? ''} />}
+          {view === 'kvm' && <KVM machines={machines} />}
+          {view === 'usage' && <Usage key={machine ?? ''} />}
           {view === 'machines' && <Machines hub={hubMe} onChanged={loadMachines} onOpen={(n) => { switchMachine(n); setView('procs') }} />}
           {view === 'widgets' && <Gallery />}
           {view === 'docker' && <Docker />}
           {view === 'alerts' && <Alerts machines={machines.map((m) => m.name)} titleOf={(n) => titleOf(n === hubMe.host ? null : n, machines, hubMe.host)} />}
           {view === 'procs' && <Top />}
           {view === 'services' && <Services />}
-          {view === 'cron' && <Cron />}
+          {view === 'jobs' && <Jobs />}
           {view === 'updates' && <Updates />}
           {view === 'disks' && <Disks />}
+          {view === 'zfs' && <ZFS />}
           {view === 'files' && <Files onOpenInTerminal={runInTerminal} />}
           {termView && <TermTabs terms={ordered} active={active} creating={creating}
             onSelect={setActive} onNew={create} onRename={rename} onClose={close} onMove={moveTab}
@@ -298,7 +309,7 @@ export default function Shell({ me: hubMe, onLoggedOut }: { me: Me; onLoggedOut:
         <span>{machine ? `${title} · ` : ''}{me.version.startsWith('v') ? me.version : 'v' + me.version}</span>
         <span>up {fmtUptime(me.uptime)}</span>
         <span className="spacer" />
-        <span className="muted">{view === 'widgets' ? 'Widgets' : view === 'docker' ? 'Docker' : view === 'alerts' ? 'Alerts' : view === 'services' ? 'Services' : view === 'cron' ? 'Cron' : view === 'updates' ? 'Updates' : view === 'disks' ? 'Disks' : view === 'space' ? 'Space' : view === 'files' ? 'Files' : view === 'procs' ? 'Processes' : view === 'machines' ? 'Machines' : view === 'users' ? 'Users' : view === 'settings' ? 'Settings' : activeTerm ? activeTerm.name : ''}</span>
+        <span className="muted">{view === 'widgets' ? 'Widgets' : view === 'docker' ? 'Docker' : view === 'alerts' ? 'Alerts' : view === 'services' ? 'Services' : view === 'jobs' ? 'Jobs' : view === 'updates' ? 'Updates' : view === 'disks' ? 'Disks' : view === 'zfs' ? 'ZFS' : view === 'usage' ? 'Usage' : view === 'files' ? 'Files' : view === 'procs' ? 'Processes' : view === 'machines' ? 'Machines' : view === 'kvm' ? 'KVM' : view === 'users' ? 'Users' : view === 'settings' ? 'Settings' : activeTerm ? activeTerm.name : ''}</span>
       </footer>
     </div>
   )

web/src/Top.tsx

@@ -250,6 +250,20 @@ export default function Top() {
 
   // the process table, and the CPU model, come from the same place the Processes page reads
   const [pt, setPt] = useState<ProcTable | null>(null)
+  const [pickedIf, setPickedIf] = useState<string | null>(null) // which NIC the net box draws
+  // Which pool each drive belongs to, asked once: a drive in a ZFS pool has no filesystem of
+  // its own, and "nothing is on it" would be a lie about eight drives holding a NAS.
+  const [pools, setPools] = useState<Record<string, string>>({})
+  useEffect(() => {
+    let live = true
+    api.storage().then((r) => {
+      if (!live) return
+      const m: Record<string, string> = {}
+      for (const d of r.devices) for (const part of d.parts) if (part.fstype === 'zfs_member' && part.label) m[d.name] = part.label
+      setPools(m)
+    }).catch(() => { /* the rows simply say less */ })
+    return () => { live = false }
+  }, [])
   const [err, setErr] = useState('')
   useEffect(() => {
     let stop = false
@@ -319,7 +333,10 @@ export default function Top() {
   const free = (mem as { free?: number }).free ?? Math.max(0, mem.total - mem.used - cached)
   const swap = s?.swap ?? pt?.swap
   const net = (s?.net ?? pt?.net ?? []).filter((n) => n.up && n.iface !== 'lo')
-  const iface = net[0]
+  // the one that has actually carried traffic, not whichever the kernel listed first — wg0
+  // and a dormant wlan should not push the real NIC out of the way. Tapping a name switches.
+  const busiest = net.length ? net.reduce((a, b) => (b.rx_total + b.tx_total > a.rx_total + a.tx_total ? b : a)) : undefined
+  const iface = net.find((n) => n.iface === pickedIf) ?? busiest
   // the disks box is this machine's own storage: a CIFS or NFS share is somebody else's disk
   const fs = (s?.fs ?? []).filter((f) => !f.remote)
   const pct = (a: number, b: number) => (b > 0 ? (a / b) * 100 : 0)
@@ -379,18 +396,26 @@ export default function Top() {
 
       <Box n={3} name="disks" className="top-disks" right="io">
         <div className="t-disks">
-          {swap && swap.total > 0 && <Disk name="swap" size={swap.total} used={swap.used} />}
+          {(s?.disks ?? []).map((d) => {
-          {fs.map((f) => {
+            // what is on this drive: the filesystems the sampler could attribute to it. A ZFS
-            // which drive is behind a mount isn't in the sample, so only the root filesystem
+            // pool or an md array spans several drives, so nothing is attributed and the row
-            // gets an IO line, against the machine's largest drive — which is the one it is on
+            // shows what is true of the hardware — how busy it is, and how hot.
-            const d = f.mount === '/' ? [...(s?.disks ?? [])].sort((a, b) => b.size - a.size)[0] : undefined
+            const on = fs.filter((f) => f.disk === d.name)
-            return <Disk key={f.mount} name={f.mount === '/' ? 'root' : f.mount.split('/').filter(Boolean).pop() ?? f.mount} size={f.size} used={f.used} io={d ? d.util_pct : undefined} />
+            const used = on.reduce((a, f) => a + f.used, 0)
+            return <Disk key={d.name} name={d.name} note={d.model} size={d.size} used={on.length ? used : undefined} io={d.util_pct} temp={d.temp} mounts={on.map((f) => f.mount)} pool={pools[d.name]} />
           })}
-          {fs.length === 0 && <div className="t-dim">no filesystems reported</div>}
+          {(s?.disks ?? []).length === 0 && <div className="t-dim">no drives reported</div>}
         </div>
       </Box>
 
-      <Box n={4} name="net" className="top-net" right={iface ? <span className="t-if">{iface.iface}</span> : undefined}>
+      <Box n={4} name="net" className="top-net" right={net.length ? (
+        <span className="t-ifs">
+          {net.map((n) => (
+            <button key={n.iface} className={'t-if' + (iface && n.iface === iface.iface ? ' on' : '')}
+              title={`${n.iface} — ${bytes(n.rx_total)} in, ${bytes(n.tx_total)} out`} onClick={() => setPickedIf(n.iface)}>{n.iface}</button>
+          ))}
+        </span>
+      ) : undefined}>
         {iface ? (
           <div className="t-net">
             <div className="t-net-graphs">
@@ -471,14 +496,29 @@ const Rows = memo(function Rows({ procs, sel, onSel, curves, empty }: { procs: P
   )
 })
 
-function Disk({ name, size, used, io }: { name: string; size: number; used: number; io?: number }) {
+// One physical drive. `used` is the bytes of the filesystems on it, left out when none can
-  const upct = size > 0 ? (used / size) * 100 : 0
+// be attributed to it — a pool member's share of a pool is not a number worth inventing.
+function Disk({ name, note, size, used, io, temp, mounts, pool }: { name: string; note?: string; size: number; used?: number; io?: number; temp?: number; mounts?: string[]; pool?: string }) {
+  const upct = used !== undefined && size > 0 ? (used / size) * 100 : 0
   return (
     <div className="t-disk">
-      <div className="t-row"><b>{name}</b><span className="spacer" />{bytes(size)}</div>
+      <div className="t-row">
+        <b>{name}</b>
+        {note && <span className="t-dim t-model" title={note}>{note}</span>}
+        <span className="spacer" />
+        {temp !== undefined && temp > 0 && <span className="t-temp">{Math.round(temp)}°C</span>}
+        {bytes(size)}
+      </div>
       {io !== undefined && <div className="t-row t-sub"><span>IO%</span><Meter pct={io} grad="io" /></div>}
+      {used !== undefined ? (
+        <>
           <div className="t-row t-sub"><span>Used:</span><span className="t-pct">{Math.round(upct)}%</span><Meter pct={upct} grad="used" /><span className="t-amt">{bytes(used, 1)}</span></div>
           <div className="t-row t-sub"><span>Free:</span><span className="t-pct">{Math.round(100 - upct)}%</span><Meter pct={100 - upct} grad="free" /><span className="t-amt">{bytes(Math.max(0, size - used), 1)}</span></div>
+          {mounts && mounts.length > 0 && <div className="t-row t-sub t-dim t-mounts" title={mounts.join('  ')}>{mounts.join(' ')}</div>}
+        </>
+      ) : (
+        <div className="t-row t-sub t-dim">{pool ? `member of ZFS pool ${pool}` : 'nothing mounted from it'}</div>
+      )}
     </div>
   )
 }

web/src/Usage.tsx

@@ -1,4 +1,5 @@
-// Space: what is using a disk, WizTree-style. Pick a disk, scan it (a parallel walk of every
+// Usage: what is using a disk, WizTree-style. (The tab was called Space until 2026-09-24;
+// the Go package and the /api/space endpoints behind it still carry the old name.) Pick a disk, scan it (a parallel walk of every
 // filesystem the disk backs, as root when the host allows), then read the result two ways:
 // a tree of folders largest-first, and a treemap of whatever folder is selected, where every
 // file is a tile sized by what it takes on disk and coloured by what kind of thing it is.
@@ -45,10 +46,10 @@ const ago = (iso?: string) => {
 const fmtDur = (a: string, b?: string) => { const ms = (b ? new Date(b).getTime() : Date.now()) - new Date(a).getTime(); return ms < 1000 ? `${ms} ms` : ms < 60000 ? `${(ms / 1000).toFixed(1)} s` : `${Math.floor(ms / 60000)} min ${Math.round((ms % 60000) / 1000)} s` }
 const parentOf = (p: string) => (p === '/' ? '/' : p.slice(0, p.lastIndexOf('/')) || '/')
 
-export default function Space() {
+export default function Usage() {
   const [targets, setTargets] = useState<SpaceTarget[] | null>(null)
   const [err, setErr] = useState('')
-  const [sel, setSel] = useHashSel('space')
+  const [sel, setSel] = useHashSel('usage')
   const [tab, setTab] = useState<'tree' | 'files'>('tree')
   const rate = useRefreshMs()
 

web/src/ZFS.tsx

@@ -0,0 +1,359 @@
+// ZFS: pools, datasets, snapshots, and a schedule for taking and expiring them.
+//
+// Split out of the Disks tab on 2026-09-24. Disks is the hardware; this is the filesystem
+// on top of it, which has enough of its own to say.
+//
+// On the schedule: `grid` has run zfSnap from root cron since before Ham Lab existed, and it
+// works. So this reads and writes the same snapshot names — `2026-09-24_02.30.01--6w`, with
+// the retention in the name — rather than inventing a second convention for the two tools to
+// disagree over. An existing cron job is detected and shown, so the page tells the truth
+// about the machine whether Ham Lab or something else is doing the work.
+import { FormEvent, useCallback, useEffect, useState } from 'react'
+import { api, ZFSExternal, ZFSInfo, ZFSOverview, ZFSPolicy, ZFSRun, ZFSSnapshot } from './api'
+import { fmtBytes } from './fmt'
+import { useRefreshMs } from './prefs'
+import { askText, askYes, Btn, Loading, Meter, Spinner, toast, useEscape } from './ui'
+
+const ago = (t?: string | number) => {
+  if (!t) return 'never'
+  const s = Math.max(0, (Date.now() - new Date(t).getTime()) / 1000)
+  return s < 60 ? 'just now' : s < 3600 ? `${Math.floor(s / 60)} min ago` : s < 86400 ? `${Math.floor(s / 3600)} h ago` : `${Math.floor(s / 86400)} d ago`
+}
+const soon = (t?: string) => {
+  if (!t) return '—'
+  const s = (new Date(t).getTime() - Date.now()) / 1000
+  return s < 60 ? 'any moment' : s < 3600 ? `in ${Math.round(s / 60)} min` : s < 86400 ? `in ${Math.round(s / 3600)} h` : `in ${Math.round(s / 86400)} d`
+}
+const WEEKDAYS = ['Sunday', 'Monday', 'Tuesday', 'Wednesday', 'Thursday', 'Friday', 'Saturday']
+const describe = (p: ZFSPolicy) => {
+  const when = p.every === 'custom' ? `on ${p.cron}`
+    : p.every === 'hourly' ? `every hour at :${(p.at.split(':')[1] ?? '00').padStart(2, '0')}`
+    : p.every === 'daily' ? `every day at ${p.at}`
+      : p.every === 'weekly' ? `every ${WEEKDAYS[p.weekday] ?? 'Sunday'} at ${p.at}`
+        : `on day ${p.day} of the month at ${p.at}`
+  return `${when} · keep ${p.keep_for}${p.recursive ? ' · recursive' : ''}`
+}
+
+// ---- pools and datasets (moved from the Disks tab) ----
+function PoolCard({ p, onChanged }: { p: ZFSInfo['pools'][number]; onChanged: () => void }) {
+  const scrubbing = /scrub in progress/i.test(p.scan ?? '')
+  const paused = /scrub paused/i.test(p.scan ?? '')
+  const [cfg, setCfg] = useState(false)
+  const act = (a: 'start' | 'pause' | 'stop') => async () => {
+    try { await api.zfsScrub(p.name, a); toast(`scrub ${a}: ok`); onChanged() } catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
+  }
+  return (
+    <section className="card">
+      <div className="sub-head">
+        <span className={'dot ' + (p.health === 'ONLINE' ? 'ok' : p.health === 'DEGRADED' ? 'warn' : 'bad')} />
+        <span className="dname">{p.name}</span>
+        <span className="badge">{p.health}</span>
+        <span className="spacer" />
+        <span className="num">{fmtBytes(p.alloc)} / {fmtBytes(p.size)} · {p.cap}%</span>
+      </div>
+      <Meter frac={p.cap / 100} cls={p.cap >= 90 ? 'bad' : p.cap >= 80 ? 'warn' : ''} />
+      <div className="props" style={{ marginTop: 8 }}>
+        <div><span className="k">Free</span><span>{fmtBytes(p.free)} · fragmentation {p.frag}% · dedup {p.dedup}</span></div>
+        {p.scan && <div><span className="k">Scan</span><span style={{ whiteSpace: 'normal' }}>{p.scan}</span></div>}
+        {p.errors && <div><span className="k">Errors</span><span className={p.errors.startsWith('No known') ? '' : 'warn-text'}>{p.errors}</span></div>}
+      </div>
+      <div className="svc-actions">
+        {!scrubbing && <Btn className="btn" confirm={`Start a scrub of ${p.name}? It reads the whole pool and can take many hours.`} onClick={act('start')}>{paused ? 'Resume scrub' : 'Scrub'}</Btn>}
+        {scrubbing && <Btn className="btn" onClick={act('pause')}>Pause scrub</Btn>}
+        {(scrubbing || paused) && <Btn className="btn" confirm="Stop the scrub?" onClick={act('stop')}>Stop scrub</Btn>}
+        <button className="btn" onClick={() => setCfg(!cfg)}>{cfg ? 'Hide vdevs' : 'Vdevs'}</button>
+      </div>
+      {cfg && <pre className="unit-file">{p.config}</pre>}
+    </section>
+  )
+}
+
+// A snapshot's name carries when it was taken and how long it lives — 2026-09-24_02.30.01--6w
+// — so the explorer can say when each one goes without asking the server anything.
+function snapMeta(full: string): { taken?: Date; expires?: Date } {
+  const snap = full.split('@')[1] ?? ''
+  const [body, ttl] = snap.split('--')
+  const m = /(\d{4})-(\d{2})-(\d{2})_(\d{2})\.(\d{2})\.(\d{2})$/.exec(body)
+  if (!m) return {}
+  const taken = new Date(+m[1], +m[2] - 1, +m[3], +m[4], +m[5], +m[6])
+  const a = /^(\d+)([mhdw])$/.exec(ttl ?? '')
+  if (!a) return { taken }
+  const ms = { m: 60e3, h: 3600e3, d: 86400e3, w: 7 * 86400e3 }[a[2]]!
+  return { taken, expires: new Date(taken.getTime() + +a[1] * ms) }
+}
+
+// One dataset in the snapshot explorer: how many it has and what they cost, opening into
+// the snapshots themselves, each of which can be taken now or deleted.
+function DatasetRow({ d, onChanged }: { d: ZFSInfo['datasets'][number]; onChanged: () => void }) {
+  const [snaps, setSnaps] = useState<ZFSSnapshot[] | null>(null)
+  const [open, setOpen] = useState(false)
+  const reload = useCallback(async () => {
+    try { setSnaps(await api.zfsSnapshots(d.name)) } catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err'); setSnaps([]) }
+  }, [d.name])
+  async function toggle() {
+    setOpen(!open)
+    if (!open && !snaps) await reload()
+  }
+  async function snapshot() {
+    const def = new Date().toISOString().slice(0, 19).replace('T', '_').replace(/:/g, '.')
+    const name = await askText(`Snapshot name for ${d.name}@`, def)
+    if (!name) return
+    try { await api.zfsSnapshot(d.name, name.trim()); toast(`${d.name}@${name.trim()} created`); await reload(); onChanged() }
+    catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
+  }
+  async function destroy(s: ZFSSnapshot) {
+    const short = s.name.split('@')[1]
+    if (!await askYes(`Delete ${s.name}? It is removed from this dataset and from every dataset beneath it that has a snapshot of the same name. This cannot be undone.`, 'Delete')) return
+    try { await api.deleteZfsSnapshot(s.name, true); toast(`${short} deleted`); await reload(); onChanged() }
+    catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
+  }
+  const depth = d.name.split('/').length - 1
+  return (
+    <>
+      <div className="ds-row" style={{ paddingLeft: 12 + depth * 16 }}>
+        <span className="ds-name" onClick={toggle}><span className="chev">{open ? '▾' : '▸'}</span>{d.name.split('/').pop()}{d.type === 'volume' && <span className="badge">zvol</span>}</span>
+        <span className="num">{d.snapshots}</span>
+        <span className="num muted">{d.used_by_snapshots ? fmtBytes(d.used_by_snapshots) : '—'}</span>
+        <span className="num muted">{fmtBytes(d.used)}</span>
+        <span className="num muted">{fmtBytes(d.refer)}</span>
+        <span className="ds-mount muted">{d.mount}</span>
+      </div>
+      {open && (
+        <div className="ds-snaps">
+          <div className="svc-actions"><Btn className="btn" onClick={snapshot}>Snapshot now</Btn><span className="muted small">{d.snapshots} snapshot{d.snapshots === 1 ? '' : 's'} · {fmtBytes(d.avail)} free on this dataset</span></div>
+          {snaps === null ? <div className="muted"><Spinner /> listing…</div> : snaps.length === 0 ? <div className="muted">No snapshots.</div> : (
+            <div className="snap-list">{snaps.slice().reverse().map((s) => {
+              const { taken, expires } = snapMeta(s.name)
+              const gone = expires && expires.getTime() < Date.now()
+              return (
+                <div key={s.name} className="snap">
+                  <span className="mono snap-name" title={s.name}>{s.name.split('@')[1]}</span>
+                  <span className="muted snap-when">{taken ? taken.toLocaleString() : new Date(s.created).toLocaleString()}</span>
+                  <span className={'muted snap-exp' + (gone ? ' warn' : '')}>{expires ? (gone ? 'expired' : `expires ${expires.toLocaleDateString()}`) : 'kept'}</span>
+                  <span className="num muted">{fmtBytes(s.used)}</span>
+                  <Btn className="btn danger snap-del" onClick={() => destroy(s)}>Delete</Btn>
+                </div>
+              )
+            })}</div>
+          )}
+        </div>
+      )}
+    </>
+  )
+}
+
+// ---- schedules ----
+
+function ExternalNotice({ x, mine, onAdopt }: { x: ExternalJob; mine: number; onAdopt: () => void }) {
+  return (
+    <div className={'notice' + (mine > 0 ? ' attn' : '')} style={{ marginBottom: 8 }}>
+      <b>{x.tool}</b> is already taking snapshots on this machine, from cron as <b>{x.user}</b>:
+      <div className="mono small" style={{ margin: '4px 0', wordBreak: 'break-all' }}>{x.schedule} &nbsp; {x.command}</div>
+      <span className="muted">in {x.source}{x.last ? ` · last ran ${ago(x.last)}` : ''}</span>
+      {mine > 0
+        ? <div className="warn" style={{ marginTop: 6 }}>A Ham Lab schedule is running as well. Both use the same names and both only remove snapshots that have passed the expiry written into them, so they will not fight — but they will take twice as many. Keep one.</div>
+        : <div className="svc-actions" style={{ marginTop: 6 }}>
+            <Btn className="btn" onClick={onAdopt}>Adopt as a schedule</Btn>
+            <span className="muted small">Copies it into a Ham Lab schedule you can edit here. Turn the old cron job off on the Jobs tab once you are happy with it.</span>
+          </div>}
+    </div>
+  )
+}
+type ExternalJob = ZFSExternal
+
+function ScheduleCard({ p, cron, next, last, onChanged, onEdit }: { p: ZFSPolicy; cron: string; next?: string; last: ZFSRun; onChanged: () => void; onEdit: () => void }) {
+  const [busy, setBusy] = useState('')
+  const run = async () => {
+    if (!await askYes(`Take a snapshot of ${p.dataset} now? Snapshots past their retention are removed at the same time, as they would be on the schedule.`, 'Run now')) return
+    setBusy('run')
+    try { const r = await api.runZfsSchedule(p.id); toast(`${r.taken}${r.pruned.length ? ` · ${r.pruned.length} expired removed` : ''}`); onChanged() }
+    catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') } finally { setBusy('') }
+  }
+  const remove = async () => {
+    if (!await askYes(`Remove the schedule "${p.name}"? The snapshots it already took are left exactly as they are.`, 'Remove')) return
+    try { await api.deleteZfsSchedule(p.id); toast('schedule removed'); onChanged() } catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
+  }
+  const toggle = async () => {
+    try { await api.saveZfsSchedule({ ...p, enabled: !p.enabled }); onChanged() } catch (e) { toast(e instanceof Error ? e.message : 'failed', 'err') }
+  }
+  return (
+    <section className="card">
+      <div className="sub-head">
+        <span className={'dot ' + (last.error ? 'bad' : p.enabled ? 'ok' : '')} />
+        <span className="dname">{p.name || p.dataset}</span>
+        {!p.enabled && <span className="badge">paused</span>}
+        <span className="spacer" />
+        <span className="muted small">{p.enabled ? soon(next) : 'not scheduled'}</span>
+      </div>
+      <div className="props" style={{ marginTop: 6 }}>
+        <div><span className="k">Dataset</span><span className="mono">{p.dataset}</span></div>
+        <div><span className="k">When</span><span>{describe(p)}</span></div>
+        <div><span className="k">Cron</span><span className="mono small" title={`in ${'/etc/cron.d/hamlab-zfs'}`}>{cron}</span></div>
+        <div><span className="k">Last run</span><span>{ago(last.at)}{last.pruned?.length ? ` · ${last.pruned.length} expired removed` : ''}</span></div>
+        {last.taken && <div><span className="k">Last snapshot</span><span className="mono small">{last.taken.split('@')[1]}</span></div>}
+        {p.keep_min > 0 && <div><span className="k">Floor</span><span>always keep the newest {p.keep_min}</span></div>}
+      </div>
+      {last.error && <div className="error" style={{ marginTop: 8 }}>{last.error}</div>}
+      <div className="svc-actions">
+        <Btn className="btn on" busy={busy === 'run'} onClick={run}>Run now</Btn>
+        <button className="btn" onClick={onEdit}>Edit</button>
+        <Btn className="btn" onClick={toggle}>{p.enabled ? 'Pause' : 'Resume'}</Btn>
+        <Btn className="btn danger" onClick={remove}>Remove</Btn>
+      </div>
+    </section>
+  )
+}
+
+const BLANK: Partial<ZFSPolicy> = { name: '', dataset: '', recursive: true, every: 'daily', at: '02:30', weekday: 0, day: 1, keep_for: '6w', keep_min: 1, enabled: true }
+
+function ScheduleDialog({ p, datasets, onClose, onSaved }: { p: Partial<ZFSPolicy>; datasets: string[]; onClose: () => void; onSaved: () => void }) {
+  const [f, setF] = useState<Partial<ZFSPolicy>>({ ...BLANK, ...p })
+  const [err, setErr] = useState('')
+  useEscape(onClose)
+  const set = (k: keyof ZFSPolicy, v: unknown) => setF((x) => ({ ...x, [k]: v }))
+  const save = async (e: FormEvent) => {
+    e.preventDefault(); setErr('')
+    try { await api.saveZfsSchedule({ ...f, name: (f.name || '').trim() || f.dataset }); onSaved() }
+    catch (e) { setErr(e instanceof Error ? e.message : 'failed') }
+  }
+  return (
+    <div className="modal-back" onClick={onClose}>
+      <form className="modal" onClick={(e) => e.stopPropagation()} onSubmit={save}>
+        <h3>{f.id ? 'Edit schedule' : 'New snapshot schedule'}</h3>
+        <label className="muted">Dataset
+          <select value={f.dataset} onChange={(e) => set('dataset', e.target.value)} required>
+            <option value="">choose…</option>
+            {datasets.map((d) => <option key={d} value={d}>{d}</option>)}
+          </select>
+        </label>
+        <label className="muted">Name — what this schedule is for
+          <input value={f.name} onChange={(e) => set('name', e.target.value)} placeholder={f.dataset || 'nightly'} maxLength={64} />
+        </label>
+        <label className="muted">How often
+          <span className="chips">
+            {(['hourly', 'daily', 'weekly', 'monthly', 'custom'] as const).map((k) => (
+              <button type="button" key={k} className={'chip' + (f.every === k ? ' on' : '')} onClick={() => set('every', k)}>{k}</button>
+            ))}
+          </span>
+        </label>
+        {f.every === 'custom' ? (
+          <label className="muted">Cron expression — minute hour day month weekday
+            <input className="mono" value={f.cron ?? ''} onChange={(e) => set('cron', e.target.value)} placeholder="0 */6 * * *" spellCheck={false} required />
+          </label>
+        ) : (
+          <label className="muted">At
+            <input type="time" value={f.at} onChange={(e) => set('at', e.target.value)} />
+          </label>
+        )}
+        {f.every === 'weekly' && (
+          <label className="muted">On
+            <select value={f.weekday} onChange={(e) => set('weekday', Number(e.target.value))}>
+              {WEEKDAYS.map((d, i) => <option key={d} value={i}>{d}</option>)}
+            </select>
+          </label>
+        )}
+        {f.every === 'monthly' && (
+          <label className="muted">Day of the month — 1 to 28, so every month has one
+            <input type="number" min={1} max={28} value={f.day} onChange={(e) => set('day', Number(e.target.value))} />
+          </label>
+        )}
+        <label className="muted">Snapshot name prefix — optional; the timestamp and retention follow it
+          <input value={f.prefix ?? ''} onChange={(e) => set('prefix', e.target.value)} placeholder="none" spellCheck={false} />
+        </label>
+        <label className="muted">Keep for — written into each snapshot's name, as zfSnap does it: 6w, 30d, 12h
+          <input value={f.keep_for} onChange={(e) => set('keep_for', e.target.value)} placeholder="6w" spellCheck={false} required />
+        </label>
+        <label className="muted">Always keep the newest — a floor, whatever their age
+          <input type="number" min={0} max={99} value={f.keep_min} onChange={(e) => set('keep_min', Number(e.target.value))} />
+        </label>
+        <label className="muted" style={{ flexDirection: 'row', alignItems: 'center', gap: 8 }}>
+          <input type="checkbox" checked={!!f.recursive} onChange={(e) => set('recursive', e.target.checked)} style={{ width: 'auto' }} />
+          Recursive — one atomic snapshot of this dataset and everything under it
+        </label>
+        <label className="muted" style={{ flexDirection: 'row', alignItems: 'center', gap: 8 }}>
+          <input type="checkbox" checked={!!f.enabled} onChange={(e) => set('enabled', e.target.checked)} style={{ width: 'auto' }} />
+          Enabled
+        </label>
+        {err && <div className="error">{err}</div>}
+        <div className="svc-actions">
+          <button type="submit" className="btn on" disabled={!f.dataset || !f.keep_for}>Save</button>
+          <span className="spacer" />
+          <button type="button" className="btn" onClick={onClose}>Cancel</button>
+        </div>
+      </form>
+    </div>
+  )
+}
+
+// adopt reads an existing cron job into the fields of a new schedule, so replacing it is a
+// matter of checking the numbers rather than retyping them. zfSnap's own flags are read
+// where they can be: -a <age> is the retention, -r is recursive.
+function adopt(x: ExternalJob, fallbackDataset: string): Partial<ZFSPolicy> {
+  const cmd = x.command
+  const keep = /-a\s+(\d+[mhdwMy])/.exec(cmd)?.[1] ?? /--\s*(\d+[mhdw])\b/.exec(cmd)?.[1] ?? '6w'
+  const f = x.schedule.trim().split(/\s+/)
+  const min = Number(f[0]), hour = Number(f[1])
+  const simple = f.length === 5 && Number.isInteger(min) && Number.isInteger(hour) && f[2] === '*' && f[3] === '*' && f[4] === '*'
+  const ds = /\b(tank[\w/-]*)\b/.exec(cmd)?.[1] ?? fallbackDataset
+  return {
+    ...BLANK, name: `${x.tool} (adopted)`, dataset: ds,
+    recursive: /\s-\w*r/.test(cmd) || true,
+    every: simple ? 'daily' : 'custom',
+    at: simple ? `${String(hour).padStart(2, '0')}:${String(min).padStart(2, '0')}` : '02:30',
+    cron: simple ? '' : x.schedule.trim(),
+    keep_for: keep, keep_min: 1, enabled: true,
+  }
+}
+
+// ---- page ----
+
+export default function ZFS() {
+  const [ov, setOv] = useState<ZFSOverview | null>(null)
+  const [err, setErr] = useState('')
+  const [editing, setEditing] = useState<Partial<ZFSPolicy> | null>(null)
+  const rate = useRefreshMs()
+
+  const load = useCallback(async () => {
+    try { setOv(await api.zfsOverview()); setErr('') } catch (e) { setErr(e instanceof Error ? e.message : 'failed') }
+  }, [])
+  useEffect(() => { load() }, [load])
+  useEffect(() => { const t = setInterval(load, Math.max(5000, rate * 10)); return () => clearInterval(t) }, [load, rate])
+
+  if (err && !ov) return <div className="dash"><div className="error">{err}</div></div>
+  if (!ov) return <div className="dash"><Loading label="Reading ZFS…" /></div>
+  const zfs: ZFSInfo = ov.zfs
+  const datasets = zfs.datasets.filter((d) => d.type !== 'volume').map((d) => d.name)
+  const mine = ov.policies.filter((x) => x.policy.enabled).length
+
+  if (!zfs.available) return <div className="dash"><div className="notice">{zfs.message || 'No ZFS on this host.'}</div></div>
+
+  return (
+    <div className="dash">
+      <div className="dash-bar"><h2 className="section">Pools</h2></div>
+      {zfs.pools.length === 0 && <div className="notice">ZFS is present but no pools are imported.</div>}
+      <div className="cards">{zfs.pools.map((p) => <PoolCard key={p.name} p={p} onChanged={load} />)}</div>
+
+      <div className="dash-bar" style={{ marginTop: 12 }}><h2 className="section">Snapshot schedules</h2><span className="spacer" />
+        <Btn className="btn on" onClick={() => setEditing({ ...BLANK, dataset: zfs.pools[0]?.name ?? '' })}>+ New schedule</Btn>
+      </div>
+      {ov.external && <ExternalNotice x={ov.external} mine={mine} onAdopt={() => setEditing(adopt(ov.external!, zfs.pools[0]?.name ?? ''))} />}
+      {ov.policies.length === 0 && !ov.external && <div className="notice">Nothing is taking snapshots on a schedule. Snapshots are written the way zfSnap writes them, with the retention in the name, so anything already here is understood.</div>}
+      <div className="cards">{ov.policies.map(({ policy, cron, next, last }) => (
+        <ScheduleCard key={policy.id} p={policy} cron={cron} next={next} last={last ?? {}} onChanged={load} onEdit={() => setEditing(policy)} />
+      ))}</div>
+
+      <div className="dash-bar" style={{ marginTop: 12 }}><h2 className="section">Snapshots</h2><span className="spacer" />
+        <span className="muted small">{zfs.datasets.reduce((a, d) => a + d.snapshots, 0)} across {zfs.datasets.length} datasets · open one to browse, take or delete</span>
+      </div>
+      <section className="card ds-table">
+        <div className="ds-row head">
+          <span>Dataset</span><span className="num">Snapshots</span><span className="num">They use</span>
+          <span className="num">Dataset uses</span><span className="num">Referenced</span><span>Mount</span>
+        </div>
+        {zfs.datasets.map((d) => <DatasetRow key={d.name} d={d} onChanged={load} />)}
+      </section>
+
+      {editing && <ScheduleDialog p={editing} datasets={datasets.length ? datasets : zfs.pools.map((p) => p.name)} onClose={() => setEditing(null)} onSaved={() => { setEditing(null); load() }} />}
+    </div>
+  )
+}

web/src/api.ts

@@ -70,7 +70,7 @@ export interface Temp { chip: string; label: string; c: number; max?: number; cr
 export interface Fan { chip: string; label: string; rpm: number }
 export interface NetIf { iface: string; up: boolean; rx_bps: number; tx_bps: number; rx_total: number; tx_total: number }
 export interface Disk { name: string; model?: string; size: number; read_bps: number; write_bps: number; util_pct: number; temp?: number }
-export interface FS { mount: string; type: string; size: number; used: number; remote?: boolean }
+export interface FS { mount: string; type: string; size: number; used: number; remote?: boolean; disk?: string }
 export interface UPS { driver: string; model: string; state: string; source: string; on_battery: boolean; in_v: number; out_v: number; battery_pct: number; runtime_min: number; load_w: number; load_pct: number; last_event?: string }
 export interface Proc { pid: number; ppid: number; name: string; cmd?: string; user: string; cpu: number; cpu_time: number; rss: number; threads: number; state: string; nice: number; started: number }
 export interface MemInfo { total: number; used: number; avail: number; free?: number; cached?: number; buffers?: number }
@@ -133,6 +133,13 @@ export interface SpaceNode { name: string; dir: boolean; size: number; apparent:
 export interface SpaceTree extends SpaceNode { path: string; children?: SpaceTree[]; more_n?: number; more_size?: number; total: number }
 export interface SpaceFile { path: string; size: number; apparent: number; mtime: number }
 
+export interface KVM { id: string; name: string; url: string; machine?: string; added: string }
+
+export interface ZFSPolicy { id: string; name: string; dataset: string; recursive: boolean; every: 'hourly' | 'daily' | 'weekly' | 'monthly' | 'custom'; at: string; weekday: number; day: number; cron?: string; keep_for: string; keep_min: number; prefix?: string; enabled: boolean }
+export interface ZFSRun { at?: string; taken?: string; pruned?: string[]; error?: string }
+export interface ZFSExternal { tool: string; schedule: string; command: string; user: string; source: string; id: string; next?: number; last?: number }
+export interface ZFSOverview { zfs: ZFSInfo; policies: { policy: ZFSPolicy; cron: string; next?: string; last: ZFSRun }[]; external?: ZFSExternal | null }
+
 export interface FileEntry { name: string; type: 'dir' | 'file' | 'link' | 'other'; size: number; mode: string; perm: string; mtime: number; user: string; group: string; target?: string; broken?: boolean }
 export interface FileInfo extends FileEntry { path: string; text: boolean }
 
@@ -179,6 +186,14 @@ export const api = {
   storage: () => req<{ devices: StorageDevice[]; zfs: ZFSInfo; smartctl: boolean }>('GET', '/api/storage'),
   smart: (dev: string, force = false) => req<SmartInfo>('GET', `/api/storage/smart/${dev}${force ? '?force=1' : ''}`),
   smartTest: (dev: string, kind: 'short' | 'long') => req<void>('POST', `/api/storage/smart/${dev}/test/${kind}`),
+  kvms: () => req<KVM[]>('GET', '/api/kvms', undefined, true),
+  addKVM: (name: string, url: string, machine: string) => req<KVM>('POST', '/api/kvms', { name, url, machine }, true),
+  removeKVM: (id: string) => req<void>('DELETE', `/api/kvms/${encodeURIComponent(id)}`, undefined, true),
+  zfsOverview: () => req<ZFSOverview>('GET', '/api/zfs'),
+  saveZfsSchedule: (p: Partial<ZFSPolicy>) => req<ZFSPolicy>('PUT', '/api/zfs/schedules', p),
+  deleteZfsSchedule: (id: string) => req<void>('DELETE', `/api/zfs/schedules/${encodeURIComponent(id)}`),
+  runZfsSchedule: (id: string) => req<{ taken: string; pruned: string[] }>('POST', `/api/zfs/schedules/${encodeURIComponent(id)}/run`),
+  deleteZfsSnapshot: (name: string, recursive: boolean) => req<void>('POST', '/api/zfs/snapshots/delete', { name, confirm: name, recursive }),
   zfsSnapshots: (dataset: string) => req<ZFSSnapshot[]>('GET', `/api/storage/zfs/snapshots?dataset=${encodeURIComponent(dataset)}`),
   zfsScrub: (pool: string, action: 'start' | 'pause' | 'stop') => req<void>('POST', `/api/storage/zfs/scrub/${encodeURIComponent(pool)}/${action}`),
   zfsSnapshot: (dataset: string, name: string) => req<void>('POST', '/api/storage/zfs/snapshot', { dataset, name }),

web/src/railorder.ts

@@ -5,8 +5,25 @@ import { useEffect, useState } from 'react'
 
 const KEY = 'hamlab.rail'
 
+// Pages that were renamed, so an order saved before it still puts the page where it was put:
+// Top became Processes, Space became Usage, Cron became Jobs. A page that was removed
+// outright (Dashboard) is not here — orderBy drops ids that match no page.
+const RENAMED: Record<string, string> = { top: 'procs', space: 'usage', cron: 'jobs' }
+
 export function getRailOrder(): string[] {
-  try { const v = JSON.parse(localStorage.getItem(KEY) ?? '[]'); return Array.isArray(v) ? v.filter((x) => typeof x === 'string') : [] } catch { return [] }
+  let raw: unknown
+  try { raw = JSON.parse(localStorage.getItem(KEY) ?? '[]') } catch { return [] }
+  if (!Array.isArray(raw)) return []
+  const seen = new Set<string>()
+  const out: string[] = []
+  for (const x of raw) {
+    if (typeof x !== 'string') continue
+    const id = RENAMED[x] ?? x
+    if (seen.has(id)) continue // a merged page keeps the earlier of the two slots
+    seen.add(id)
+    out.push(id)
+  }
+  return out
 }
 
 export function setRailOrder(ids: string[], save = true) {

web/src/sampleAdapter.ts

@@ -51,7 +51,7 @@ export function sampleOf(v: Values, cat: Catalog): Sample {
     fans: hw('fan.').map(({ id, s }) => ({ chip: s.labels!.chip, label: s.labels!.label, rpm: num(v, id) })),
     net: [...groups(v, cat, 'net.', 'iface')].map(([iface, p]) => ({ iface, up: v[p + '.up'] === true, rx_bps: num(v, p + '.rx'), tx_bps: num(v, p + '.tx'), rx_total: num(v, p + '.rx_total'), tx_total: num(v, p + '.tx_total') })),
     disks: [...groups(v, cat, 'disk.', 'device')].map(([name, p]) => ({ name, model: cat.get(p + '.size')?.labels?.model || undefined, size: num(v, p + '.size'), read_bps: num(v, p + '.read'), write_bps: num(v, p + '.write'), util_pct: num(v, p + '.util'), temp: num(v, p + '.temp') || undefined })),
-    fs: [...groups(v, cat, 'fs.', 'mount')].map(([mount, p]) => ({ mount, type: cat.get(p + '.size')?.labels?.fstype ?? '', size: num(v, p + '.size'), used: num(v, p + '.used'), remote: cat.get(p + '.size')?.labels?.remote === '1' })),
+    fs: [...groups(v, cat, 'fs.', 'mount')].map(([mount, p]) => ({ mount, type: cat.get(p + '.size')?.labels?.fstype ?? '', size: num(v, p + '.size'), used: num(v, p + '.used'), remote: cat.get(p + '.size')?.labels?.remote === '1', disk: cat.get(p + '.size')?.labels?.disk })),
     ups: v['ups.battery'] === undefined ? null : {
       driver: cat.get('ups.last_event')?.labels?.driver ?? '', model: cat.get('ups.battery')?.labels?.model ?? '', state: str(v, 'ups.state'), source: str(v, 'ups.source'),
       on_battery: v['ups.on_battery'] === true, in_v: num(v, 'ups.in_v'), out_v: num(v, 'ups.out_v'), battery_pct: num(v, 'ups.battery'), runtime_min: num(v, 'ups.runtime'),

web/src/styles.css

@@ -285,14 +285,21 @@ h2.section { margin: 0; font-size: 13px; font-weight: 600; color: var(--muted);
    the fallback above stays a fallback. */
 .disk-cards { grid-template-columns: repeat(auto-fill, minmax(min(420px, 100%), 1fr)); }
 .datasets { padding: 0; overflow-x: auto; }
-.ds-row { display: grid; grid-template-columns: minmax(160px, 1fr) 80px 80px 80px 50px 50px minmax(120px, 1fr); gap: 8px; align-items: center; padding: 6px 12px; border-bottom: 1px solid var(--border); font-size: 13px; min-width: 700px; } /* the columns and gaps add up to 692: at 640 the row ended before its own cells did */
+.ds-row { display: grid; grid-template-columns: minmax(160px, 1fr) 80px 90px 90px 90px minmax(120px, 1fr); gap: 8px; align-items: center; padding: 6px 12px; border-bottom: 1px solid var(--border); font-size: 13px; min-width: 700px; }
 .ds-row.head { color: var(--muted); font-size: 11px; text-transform: uppercase; letter-spacing: .5px; }
 .ds-row .num { text-align: right; }
 .ds-name { cursor: pointer; display: flex; align-items: center; gap: 6px; white-space: nowrap; overflow: hidden; text-overflow: ellipsis; }
 .ds-mount { overflow: hidden; text-overflow: ellipsis; white-space: nowrap; font-family: ui-monospace, monospace; font-size: 12px; }
 .ds-snaps { padding: 8px 12px 10px 40px; border-bottom: 1px solid var(--border); background: var(--bg); display: flex; flex-direction: column; gap: 6px; }
 .snap-list { display: flex; flex-direction: column; gap: 2px; max-height: 40vh; overflow-y: auto; }
-.snap { display: flex; justify-content: space-between; gap: 12px; font-size: 12px; }
+.snap { display: grid; grid-template-columns: minmax(150px, 1fr) 150px 120px 70px auto; gap: 10px; align-items: center; font-size: 12px; padding: 2px 0; border-bottom: 1px solid var(--border); }
+.snap .snap-name { overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }
+.snap .snap-when, .snap .snap-exp { white-space: nowrap; }
+.snap .snap-exp.warn { color: var(--s2); }
+.snap .num { text-align: right; }
+.snap .snap-del { padding: 0 6px; font-size: 11px; opacity: .55; }
+.snap:hover .snap-del { opacity: 1; }
+@media (max-width: 800px) { .snap { grid-template-columns: minmax(120px, 1fr) auto auto; } .snap .snap-when, .snap .snap-exp { display: none; } }
 
 /* files */
 .files { position: absolute; inset: 0; display: flex; flex-direction: column; background: var(--bg); }
@@ -720,3 +727,26 @@ pre.diff .add { color: var(--ok); background: color-mix(in srgb, var(--ok) 12%,
   .space-row.file { grid-template-columns: minmax(120px, 1fr) 72px 80px; }
   .space-split { padding: 8px; } .space-tree { flex: 1 1 45%; } .space-map { flex: 1 1 55%; }
 }
+
+/* the Processes tab's drive rows and NIC switcher */
+.t-disk .t-model { overflow: hidden; text-overflow: ellipsis; white-space: nowrap; max-width: 14ch; margin-left: 6px; }
+.t-disk .t-temp { margin-right: 8px; color: var(--t-blue, var(--accent2)); }
+.t-disk .t-mounts { overflow: hidden; text-overflow: ellipsis; white-space: nowrap; }
+.t-ifs { display: inline-flex; gap: 4px; flex-wrap: wrap; justify-content: flex-end; }
+.t-ifs .t-if { padding: 0 4px; border-radius: 3px; color: var(--muted); }
+.t-ifs .t-if:hover { color: var(--fg); background: var(--bg3); }
+.t-ifs .t-if.on { color: var(--fg); box-shadow: inset 0 -2px 0 var(--accent); }
+
+/* ZFS: the schedule dialog's controls match the rest of the modal forms */
+.modal select, .modal input[type=time], .modal input[type=number] { padding: 9px 10px; background: var(--bg); border: 1px solid var(--border); border-radius: 6px; color: var(--fg); font-size: 16px; }
+.modal .chips { margin-top: 2px; }
+
+/* ---- KVM: a device's own web page, framed ---- */
+.kvm { position: absolute; inset: 0; display: flex; flex-direction: column; background: var(--bg); min-height: 0; }
+.kvm-bar { display: flex; align-items: center; gap: 8px; padding: 8px 12px; border-bottom: 1px solid var(--border); background: var(--bg2); flex-wrap: wrap; }
+.kvm-bar .kvm-url { overflow: hidden; text-overflow: ellipsis; white-space: nowrap; max-width: 28ch; }
+.kvm-frame { flex: 1; min-height: 0; background: #000; position: relative; }
+.kvm-frame iframe { width: 100%; height: 100%; border: 0; display: block; background: #000; }
+.kvm-frame.full { background: #000; }
+.kvm-hint { padding: 6px 12px; border-top: 1px solid var(--border); }
+@media (max-width: 800px) { .kvm-bar .kvm-url { display: none; } }

web/src/ui.tsx

@@ -14,6 +14,12 @@ export function CloseBtn({ onClick, title = 'Close' }: { onClick: () => void; ti
   )
 }
 
+// A filled bar. Red past 95%, amber past 85%, unless the caller says otherwise.
+export function Meter({ frac, cls }: { frac: number; cls?: string }) {
+  const c = cls ?? (frac >= 0.95 ? 'bad' : frac >= 0.85 ? 'warn' : '')
+  return <div className="meter"><div className={'fill ' + c} style={{ width: `${Math.min(100, Math.max(0, frac * 100))}%` }} /></div>
+}
+
 export function Loading({ label = 'Loading…' }: { label?: string }) {
   return <div className="center muted loading"><Spinner size={22} /><div>{label}</div></div>
 }