diff --git a/tailscale-optimize-router.sh b/tailscale-optimize-router.sh
deleted file mode 100755
index 7a17681..0000000
--- a/tailscale-optimize-router.sh
+++ /dev/null
@@ -1,4 +0,0 @@
-#!/bin/sh
-# run from root's crontab
-/usr/sbin/ethtool -K vmbr0 rx-udp-gro-forwarding on rx-gro-list off
-/usr/sbin/ethtool -K enp4s1f12np12 rx-udp-gro-forwarding on rx-gro-list off
diff --git a/unused/tailscale-optimize-router.sh b/unused/tailscale-optimize-router.sh
new file mode 100755
index 0000000..1e77fd1
--- /dev/null
+++ b/unused/tailscale-optimize-router.sh
@@ -0,0 +1,24 @@
+#!/bin/sh
+# RETIRED. Kept for reference only.
+#
+# Tailscale's UDP GRO forwarding tuning belongs on the machine running
+# tailscaled. Here that is VM 100 (ts-subnet-rtr), a KVM guest with its own
+# kernel and its own netdevs, so setting it on the Proxmox host cannot affect
+# it. Apply it inside that VM instead, on its default-route interface:
+#
+#   NETDEV=$(ip -o route get 8.8.8.8 | cut -f 5 -d " ")
+#   ethtool -K "$NETDEV" rx-udp-gro-forwarding on rx-gro-list off
+#
+# Host-side tuning would have been the right lever only if tailscaled ran in an
+# LXC container, which shares the host kernel.
+#
+# The interfaces below were wrong regardless. VM 100 sits on vmbr7, backed by
+# enp4s0f7np7; vmbr0 is backed by enp4s0f0np0 and carries different guests,
+# and enp4s1f12np12 is the host's own routing NIC, not a bridge port.
+#
+# It was also never in service: the header said "run from root's crontab", but
+# root has no crontab.
+
+# run from root's crontab
+/usr/sbin/ethtool -K vmbr0 rx-udp-gro-forwarding on rx-gro-list off
+/usr/sbin/ethtool -K enp4s1f12np12 rx-udp-gro-forwarding on rx-gro-list off