more updates

Author: abhinavnagaraj

scripts/kube-join.sh

@@ -15,8 +15,8 @@ proxy_http=$4
 proxy_https=$5
 proxy_no=$6
 
-export PATH="$PATH:$root_path/usr/bin"
-export PATH="$PATH:$root_path/usr/local/bin"
+export PATH="$PATH:$(join_path "$root_path" "usr/bin")"
+export PATH="$PATH:$(join_path "$root_path" "usr/local/bin")"
 
 KUBE_VIP_LOC="/etc/kubernetes/manifests/kube-vip.yaml"
 
@@ -26,22 +26,49 @@ MAX_RETRIES=3
 ETCD_HEALTH_TIMEOUT=60
 RETRY_DELAY=30
 
-restart_containerd() {
-  if systemctl cat spectro-containerd >/dev/null 2<&1; then
-    systemctl restart spectro-containerd
-  fi
+# Function to detect containerd socket path
+detect_containerd_socket() {
+  # Common containerd socket paths across different distributions
+  local possible_paths=(
+    "/var/run/containerd/containerd.sock"
+    "/run/containerd/containerd.sock"
+    "/var/run/dockershim.sock"
+    "/run/dockershim.sock"
+  )
+
+  # Check common socket paths
+  for path in "${possible_paths[@]}"; do
+    if [ -S "$path" ]; then
+      echo "unix://$path"
+      return 0
+    fi
+  done
 
-  if systemctl cat containerd >/dev/null 2<&1; then
-    systemctl restart containerd
-  fi
+  # If nothing found, return default
+  echo "unix:///var/run/containerd/containerd.sock"
+}
+
+# Function to join paths properly, avoiding double slashes
+join_path() {
+  local base_path="$1"
+  local relative_path="$2"
+
+  # Remove trailing slash from base_path and leading slash from relative_path
+  base_path="${base_path%/}"
+  relative_path="${relative_path#/}"
+
+  # Join them with a single slash
+  echo "${base_path}/${relative_path}"
 }
 
 # Function to wait for static pods to be created and running
 wait_for_static_pods() {
   local timeout=$1
   local start_time=$(date +%s)
+  local containerd_socket=$(detect_containerd_socket)
 
   echo "Waiting for static pods to be created and running..."
+  echo "Using containerd socket: $containerd_socket"
 
   # Check if crictl is available
   if ! command -v crictl >/dev/null 2>&1; then
@@ -63,8 +90,8 @@ wait_for_static_pods() {
     # Check if all required static pod manifests exist
     if [ -f /etc/kubernetes/manifests/etcd.yaml ] && [ -f /etc/kubernetes/manifests/kube-apiserver.yaml ] && [ -f /etc/kubernetes/manifests/kube-controller-manager.yaml ] && [ -f /etc/kubernetes/manifests/kube-scheduler.yaml ]; then
       # Check if etcd container is running
-      local etcd_container_id=$(crictl ps --name etcd --quiet 2>/dev/null)
-      if [ -n "$etcd_container_id" ] && crictl ps --id "$etcd_container_id" --format table | grep -q "Running"; then
+      local etcd_container_id=$(crictl --runtime-endpoint="$containerd_socket" ps --name etcd --quiet 2>/dev/null)
+      if [ -n "$etcd_container_id" ] && crictl --runtime-endpoint="$containerd_socket" ps --id "$etcd_container_id" | grep -q "Running"; then
         echo "Static pods are running, waiting additional time for etcd to stabilize..."
         sleep 10
         return 0
@@ -81,6 +108,7 @@ wait_for_static_pods() {
 check_etcd_health_if_available() {
   local timeout=$1
   local start_time=$(date +%s)
+  local containerd_socket=$(detect_containerd_socket)
 
   # Check if etcdctl is available
   if ! command -v etcdctl >/dev/null 2>&1; then
@@ -95,11 +123,12 @@ check_etcd_health_if_available() {
   fi
 
   echo "Checking etcd health with etcdctl..."
+  echo "Using containerd socket: $containerd_socket"
 
   while [ $(($(date +%s) - start_time)) -lt $timeout ]; do
     # Check if etcd container is running
-    local etcd_container_id=$(crictl ps --name etcd --quiet 2>/dev/null)
-    if [ -n "$etcd_container_id" ] && crictl ps --id "$etcd_container_id" --format table | grep -q "Running"; then
+    local etcd_container_id=$(crictl --runtime-endpoint="$containerd_socket" ps --name etcd --quiet 2>/dev/null)
+    if [ -n "$etcd_container_id" ] && crictl --runtime-endpoint="$containerd_socket" ps --id "$etcd_container_id" | grep -q "Running"; then
       # Try to connect to local etcd directly
       if timeout 10s etcdctl endpoint health --endpoints=https://127.0.0.1:2379 --cacert=/etc/kubernetes/pki/etcd/ca.crt --cert=/etc/kubernetes/pki/etcd/server.crt --key=/etc/kubernetes/pki/etcd/server.key >/dev/null 2>&1; then
         echo "etcd is healthy"
@@ -130,32 +159,38 @@ cleanup_cluster_state() {
   fi
 }
 
+restart_containerd() {
+  if systemctl cat containerd >/dev/null 2<&1; then
+    systemctl restart containerd
+  fi
+}
+
 do_kubeadm_reset() {
-  if [ -S /run/spectro/containerd/containerd.sock ]; then
-    kubeadm reset -f --cri-socket unix:///run/spectro/containerd/containerd.sock --cleanup-tmp-dir
+  local containerd_socket=$(detect_containerd_socket)
+
+  # Check if the detected socket exists
+  local socket_path="${containerd_socket#unix://}"
+  if [ -S "$socket_path" ]; then
+    kubeadm reset -f --cri-socket "$containerd_socket" --cleanup-tmp-dir
   else
     kubeadm reset -f --cleanup-tmp-dir
   fi
   iptables -F && iptables -t nat -F && iptables -t mangle -F && iptables -X && rm -rf /etc/kubernetes/etcd /etc/kubernetes/manifests /etc/kubernetes/pki
-  rm -rf "$root_path"/etc/cni/net.d
-  if [ -f /run/systemd/system/etc-cni-net.d.mount ]; then
-    mkdir -p "$root_path"/etc/cni/net.d
-    systemctl restart etc-cni-net.d.mount
-  fi
+  rm -rf "$(join_path "$root_path" "etc/cni/net.d")"
   systemctl daemon-reload
   restart_containerd
 }
 
 backup_kube_vip_manifest_if_present() {
   if [ -f "$KUBE_VIP_LOC" ] && [ "$NODE_ROLE" != "worker" ]; then
-    cp $KUBE_VIP_LOC "$root_path"/opt/kubeadm/kube-vip.yaml
+    cp $KUBE_VIP_LOC "$(join_path "$root_path" "opt/kubeadm/kube-vip.yaml")"
   fi
 }
 
 restore_kube_vip_manifest_after_reset() {
-  if [ -f "$root_path/opt/kubeadm/kube-vip.yaml" ] && [ "$NODE_ROLE" != "worker" ]; then
-    mkdir -p "$root_path"/etc/kubernetes/manifests
-    cp "$root_path"/opt/kubeadm/kube-vip.yaml $KUBE_VIP_LOC
+  if [ -f "$(join_path "$root_path" "opt/kubeadm/kube-vip.yaml")" ] && [ "$NODE_ROLE" != "worker" ]; then
+    mkdir -p "$(join_path "$root_path" "etc/kubernetes/manifests")"
+    cp "$(join_path "$root_path" "opt/kubeadm/kube-vip.yaml")" $KUBE_VIP_LOC
   fi
 }
 
@@ -167,9 +202,9 @@ attempt_kubeadm_join() {
   echo "Attempting kubeadm join (attempt $attempt/$max_retries)"
 
   if [ "$PROXY_CONFIGURED" = true ]; then
-    HTTP_PROXY=$proxy_http http_proxy=$proxy_http HTTPS_PROXY=$proxy_https https_proxy=$proxy_https NO_PROXY=$proxy_no no_proxy=$proxy_no kubeadm join --config "$root_path"/opt/kubeadm/kubeadm.yaml --ignore-preflight-errors=DirAvailable--etc-kubernetes-manifests -v=5
+    HTTP_PROXY=$proxy_http http_proxy=$proxy_http HTTPS_PROXY=$proxy_https https_proxy=$proxy_https NO_PROXY=$proxy_no no_proxy=$proxy_no kubeadm join --config "$(join_path "$root_path" "opt/kubeadm/kubeadm.yaml")" --ignore-preflight-errors=DirAvailable--etc-kubernetes-manifests -v=5
   else
-    kubeadm join --config "$root_path"/opt/kubeadm/kubeadm.yaml --ignore-preflight-errors=DirAvailable--etc-kubernetes-manifests -v=5
+    kubeadm join --config "$(join_path "$root_path" "opt/kubeadm/kubeadm.yaml")" --ignore-preflight-errors=DirAvailable--etc-kubernetes-manifests -v=5
   fi
 }