You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
fix: verify rescue in stateInstallTalos before installing
Root cause of repeated installer failures: SSH reachable but server
running Debian (not rescue). stateCheckRescueActive had rescue
detection, but stateInstallTalos didn't — it trusted SSH=rescue.
Fix: Before installing, verify hostname=rescue or /etc/hetzner-build.
If not rescue, fix EFI boot order (delete non-PXE entries), reboot,
re-activate rescue. Server will PXE boot into rescue on next cycle.
This affected: node3, node5, node9, node1, storage-7 — all servers
with pre-existing OS (Debian from cephadm or old Talos).
// Fix EFI boot order — delete non-PXE entries so next boot goes to PXE rescue
537
+
fixClient:=sshrescue.New(serverIP, privateKey)
538
+
iffixErr:=fixClient.Connect(); fixErr==nil {
539
+
_, _=fixClient.Run(`
540
+
if command -v efibootmgr > /dev/null 2>&1; then
541
+
mount -o remount,rw /sys/firmware/efi/efivars 2>/dev/null || \
542
+
mount -t efivarfs efivarfs /sys/firmware/efi/efivars 2>/dev/null || true
543
+
for entry in $(efibootmgr 2>/dev/null | grep '^Boot[0-9A-Fa-f]' | grep -iv 'pxe\|network\|ipv4\|ipv6' | grep -o '^Boot[0-9A-Fa-f]*' | sed 's/Boot//'); do
0 commit comments