fix: auto-detect SSH login user for node operations
All checks were successful
Terraform Plan / Terraform Plan (push) Successful in 18s

This commit is contained in:
2026-02-28 19:25:48 +00:00
parent 6105a314b7
commit 244887e9c2
2 changed files with 43 additions and 3 deletions

View File

@@ -17,6 +17,7 @@ source "$INVENTORY_FILE"
SSH_USER="${SSH_USER:-micqdf}"
SSH_KEY_PATH="${SSH_KEY_PATH:-$HOME/.ssh/id_ed25519}"
SSH_OPTS="${SSH_OPTS:--o BatchMode=yes -o IdentitiesOnly=yes -o StrictHostKeyChecking=accept-new -i $SSH_KEY_PATH}"
SSH_USER_CANDIDATES="${SSH_USER_CANDIDATES:-$SSH_USER root}"
declare -A NODE_IPS=()
declare -a CP_NAMES=()
@@ -78,7 +79,23 @@ populate_nodes() {
remote() {
local host_ip="$1"
local cmd="$2"
ssh $SSH_OPTS "$SSH_USER@$host_ip" "$cmd"
ssh $SSH_OPTS "$ACTIVE_SSH_USER@$host_ip" "$cmd"
}
detect_ssh_user() {
local probe_ip="$1"
local candidate
for candidate in $SSH_USER_CANDIDATES; do
if ssh $SSH_OPTS "$candidate@$probe_ip" "true" >/dev/null 2>&1; then
ACTIVE_SSH_USER="$candidate"
echo "==> Using SSH user '$ACTIVE_SSH_USER'"
return 0
fi
done
echo "Unable to authenticate to $probe_ip with candidates: $SSH_USER_CANDIDATES"
return 1
}
prepare_known_hosts() {
@@ -109,7 +126,7 @@ rebuild_node() {
echo "==> Rebuilding $node_name on $node_ip"
nixos-rebuild switch \
--flake "$FLAKE_DIR#$node_name" \
--target-host "$SSH_USER@$node_ip" \
--target-host "$ACTIVE_SSH_USER@$node_ip" \
--use-remote-sudo
}
@@ -122,6 +139,8 @@ if [ -z "${NODE_IPS[$PRIMARY_CONTROL_PLANE]:-}" ]; then
PRIMARY_CONTROL_PLANE="${CP_NAMES[0]}"
fi
PRIMARY_CP_IP="${NODE_IPS[$PRIMARY_CONTROL_PLANE]}"
ACTIVE_SSH_USER="$SSH_USER"
detect_ssh_user "$PRIMARY_CP_IP"
for node in "${CP_NAMES[@]}"; do
rebuild_node "$node" "${NODE_IPS[$node]}"