121 lines
4.4 KiB
Bash
Executable File
121 lines
4.4 KiB
Bash
Executable File
#!/bin/sh
|
||
|
||
ERR_LOG=/var/log/apinto/error.log
|
||
echo_info() {
|
||
echo "[$(date "+%Y-%m-%d %H:%M:%S")] [INFO] $1" >> $ERR_LOG
|
||
}
|
||
|
||
echo_error() {
|
||
echo "[$(date "+%Y-%m-%d %H:%M:%S")] [ERROR] $1" >> $ERR_LOG
|
||
}
|
||
|
||
# 检查环境变量
|
||
if [ -z "$SERVICE" ] || [ -z "$NAMESPACE" ]; then
|
||
echo_error "Environment variables SERVICE and NAMESPACE must be set."
|
||
exit 1
|
||
fi
|
||
|
||
# 解析当前 Pod 的序号
|
||
CURRENT_INDEX=${HOSTNAME##*-}
|
||
BASE_NAME=${HOSTNAME%-*}
|
||
MAX_ATTEMPTS=60 # 最大尝试次数（节点等待），约1分钟
|
||
RETRY_INTERVAL=5 # 重试间隔，单位秒
|
||
MAX_JOIN_RETRIES=12 # 加入集群失败的最大重试次数
|
||
MAX_POD_INDEX=${MAX_POD_INDEX:-10} # 默认检查最多10个Pod，可通过环境变量配置
|
||
|
||
# 等待本地服务启动
|
||
attempt=0
|
||
until curl --max-time 5 --silent --fail http://127.0.0.1:9400 || [ $attempt -ge $MAX_ATTEMPTS ]; do
|
||
echo_info "Waiting for localhost to be ready... Attempt $attempt"
|
||
sleep 1
|
||
attempt=$((attempt + 1))
|
||
done
|
||
if [ $attempt -ge $MAX_ATTEMPTS ]; then
|
||
echo_error "Timeout waiting for localhost to be ready after $MAX_ATTEMPTS attempts."
|
||
exit 1
|
||
fi
|
||
|
||
# 检查是否成功加入集群
|
||
check_cluster_join() {
|
||
info_output=$(./apinto info 2>&1)
|
||
peer_count=$(echo "$info_output" | grep -c -- "--Peer")
|
||
if [ "$peer_count" -ge 2 ]; then
|
||
echo_info "Successfully joined the cluster with $peer_count peers. Cluster info: $info_output"
|
||
return 0
|
||
else
|
||
echo_info "Failed to join the cluster. Only $peer_count peer(s) found. Info: $info_output"
|
||
return 1
|
||
fi
|
||
}
|
||
|
||
# 尝试加入集群，带重试
|
||
try_join_cluster() {
|
||
local target_addr=$1
|
||
local join_retries=0
|
||
while [ $join_retries -lt $MAX_JOIN_RETRIES ]; do
|
||
echo_info "Attempting to join cluster via $target_addr (Retry $join_retries/$MAX_JOIN_RETRIES)..."
|
||
join_output=$(./apinto join -addr "$target_addr" 2>&1)
|
||
if [ $? -eq 0 ]; then
|
||
if check_cluster_join; then
|
||
return 0
|
||
else
|
||
echo_info "Join via $target_addr executed but cluster validation failed.Details: $join_output"
|
||
fi
|
||
else
|
||
echo_info "Join via $target_addr failed. Details: $join_output"
|
||
fi
|
||
join_retries=$((join_retries + 1))
|
||
if [ $join_retries -lt $MAX_JOIN_RETRIES ]; then
|
||
echo_info "Retrying join in $RETRY_INTERVAL seconds..."
|
||
sleep $RETRY_INTERVAL
|
||
fi
|
||
done
|
||
echo_error "Failed to join cluster via $target_addr after $MAX_JOIN_RETRIES retries."
|
||
return 1
|
||
}
|
||
|
||
if [ "$CURRENT_INDEX" -eq 0 ]; then
|
||
# apinto-0: 检查其他 Pod 是否在运行
|
||
echo_info "This is $HOSTNAME. Checking if other nodes are running..."
|
||
for i in $(seq 1 "$MAX_POD_INDEX"); do
|
||
OTHER_POD="${BASE_NAME}-${i}.${SERVICE}.${NAMESPACE}.svc.cluster.local"
|
||
attempt=0
|
||
while [ $attempt -lt $MAX_ATTEMPTS ]; do
|
||
if curl --max-time 5 --silent --fail http://${OTHER_POD}:9401; then
|
||
echo_info "Found a running node: ${OTHER_POD}."
|
||
if try_join_cluster "${OTHER_POD}:9401"; then
|
||
exit 0
|
||
fi
|
||
echo_info "Failed to join via ${OTHER_POD}. Trying next node..."
|
||
break
|
||
else
|
||
echo_info "${OTHER_POD} is not ready yet. Retrying in $RETRY_INTERVAL seconds..."
|
||
fi
|
||
sleep $RETRY_INTERVAL
|
||
attempt=$((attempt + 1))
|
||
done
|
||
echo_info "Timeout waiting for ${OTHER_POD} after $MAX_ATTEMPTS attempts."
|
||
done
|
||
echo_info "No other nodes are available or joinable. Starting as the first node."
|
||
else
|
||
# 非 apinto-0 的 Pod，加入 apinto-0
|
||
LEADER_POD="${BASE_NAME}-0.${SERVICE}.${NAMESPACE}.svc.cluster.local"
|
||
echo_info "This is $HOSTNAME. Waiting for $LEADER_POD to be ready..."
|
||
attempt=0
|
||
until curl --max-time 5 --silent --fail http://$LEADER_POD:9401 || [ $attempt -ge $MAX_ATTEMPTS ]; do
|
||
echo_info "Waiting for $LEADER_POD to be ready... Attempt $attempt"
|
||
sleep 1
|
||
attempt=$((attempt + 1))
|
||
done
|
||
if [ $attempt -ge $MAX_ATTEMPTS ]; then
|
||
echo_error "Timeout waiting for $LEADER_POD to be ready after $MAX_ATTEMPTS attempts."
|
||
exit 1
|
||
fi
|
||
echo_info "$LEADER_POD is ready."
|
||
if try_join_cluster "$LEADER_POD:9401"; then
|
||
exit 0
|
||
else
|
||
echo_error "Failed to join cluster via $LEADER_POD after $MAX_JOIN_RETRIES retries."
|
||
exit 1
|
||
fi
|
||
fi |