hive TLS: reload the trust bundle's unit on a switch, name the missing root when it's empty

A switch that changed `hive-tls-ca` stopped and started it, and the stop
propagated through `Requires=` to the store's container (sealed again under
the shamir seal), every container binding the bundle, and every
`<name>-ca-bundle` unit with its consumers. `reloadIfChanged` with the same
script as `ExecReload` rewrites the bundle in place and restarts none of them.
`Requires=` stays, so a failed assembly still keeps them from starting.

With `swarm-services-root.pem` missing, `hive-tls-ca` writes an empty bundle
and every `<name>-ca-bundle` unit fails. Its log line now names the root's
path. docs/swarm/ca.md describes both.

docs/swarm/ca.md

@@ -77,8 +77,21 @@ issues it.
 `hive-tls-ca.service` writes `<deploy.hive-controller.tls.stateDir>/trust-bundle.pem`,
 the file every consumer anchors on. It holds one certificate, a copy of
 `swarm-services-root.pem`: the store's `pki` root, which the gateway leaf and
-the swarm-services leaf both chain to. It's the only root the swarm has. Until
+the swarm-services leaf both chain to. It's the only root the swarm has.
-the first fetch of that root lands, the bundle is empty.
+
+**Without the store's root, the bundle is empty and its readers go red.**
+While `<deploy.hive-controller.tls.stateDir>/swarm-services-root.pem` is
+missing or empty, as on a first boot before the store answers,
+`hive-tls-ca.service` writes an empty bundle and logs that file's path. Every
+`<name>-ca-bundle` unit then fails with `<bundle> contributed no certificate
+to the bundle`, and the consumers that `Requires=` it don't start.
+`swarm-services-cert.service` restarts `hive-tls-ca.service` when it writes a
+new root.
+
+A deploy that changes `hive-tls-ca.service` reloads it: the same script
+rewrites the bundle in place, and nothing that `Requires=` the unit restarts,
+including the store's container. A `<name>-ca-bundle` unit keeps the copy it
+made when it started.
 
 ⚠️ **A consumer reads its trust when it starts.** An agent or service
 container that last started before the store's root reached the bundle

nix/host-modules/hive-tls.nix

@@ -362,6 +362,17 @@ in
       # "$svcroot" ]` guard is what makes that state legal rather than fatal,
       # and `swarm-services-cert` restarts this unit once the root it wrote
       # is new — so the bundle converges instead of being a boot behind.
+      #
+      # ⚠️ A switch that changes this unit reloads it, running the same
+      # script, rather than stopping and starting it. A stop propagates to
+      # every unit that `Requires=` this one: the store's container, which
+      # comes back sealed under the `shamir` seal, every container that
+      # binds the bundle, and every `<name>-ca-bundle` unit with its
+      # consumers. A reload rewrites the bundle in place and touches none of
+      # them, so a `<name>-ca-bundle` unit keeps the copy it made at start
+      # until it restarts.
+      reloadIfChanged = true;
+      reload = config.systemd.services.hive-tls-ca.script;
       serviceConfig = {
         Type = "oneshot";
         RemainAfterExit = true;
@@ -387,7 +398,7 @@ in
         if [ -s "$svcroot" ]; then
           cat "$svcroot" > "$bundle"
         else
-          echo "no trust anchor on disk yet — $bundle is empty until the store's root is fetched" >&2
+          echo "the store's root $svcroot is missing or empty, so $bundle is empty and every *-ca-bundle unit fails until swarm-services-cert fetches it" >&2
           : > "$bundle"
         fi
         chmod 0644 "$bundle"

nix/module-eval/hive-tls.nix

@@ -281,6 +281,32 @@ let
         && !(lib.hasInfix "anchors+=(" caUnit.script)
         && !(lib.hasInfix "ca.pem" caUnit.script);
     }
+    {
+      # A reload does not propagate through `Requires=`; a stop and start
+      # restarts the store's container, which comes back sealed. The control
+      # is that container still requiring the unit, so a failed assembly
+      # still keeps it from starting.
+      name = "a switch that changes the trust bundle's unit reloads it with the same script, so the store's container is not restarted";
+      ok =
+        let
+          text = allLocal.systemd.units."hive-tls-ca.service".text;
+          store = units."container@swarm-bao";
+        in
+        caUnit.reloadIfChanged
+        && caUnit.reload == caUnit.script
+        && lib.hasInfix "X-ReloadIfChanged=true" text
+        && lib.hasInfix "ExecReload=" text
+        && lib.elem "hive-tls-ca.service" store.requires
+        && lib.elem "hive-tls-ca.service" store.after;
+    }
+    {
+      # A missing root is the one case where the bundle is empty, and every
+      # `<name>-ca-bundle` unit fails on it; the log names the file.
+      name = "an empty trust bundle is logged naming the store's root it is copied from";
+      ok = lib.hasInfix ''echo "the store's root $svcroot is missing or empty'' (
+        lib.last (lib.splitString ''if [ -s "$svcroot" ]; then'' caUnit.script)
+      );
+    }
     {
       # The control is the all-local host, which defines none of them.
       name = "a definition of a removed swarm.ca option fails evaluation, naming the option and the store's root";