Files
ai-security-xdr/haobang-security-dm/upgrade-doc/v3/kafka/02-create-large-log-topics-docker.sh
T
王绍宏 0e9e6656db feat(syslog): 优化大日志通道并增加异常连续消费兜底
- 128 KiB 及以上日志默认进入大日志 Topic
- InfluxDB 原文及达梦大字段超过 256 KiB 时截断保存
- 截断内容追加 HB_XDR_TRUNCATED 标记
- InfluxDB 写入失败时继续执行规则解析和达梦写入
- 未知或系统性异常优先写入普通/大日志 DLT
- DLT 写入失败时记录特殊告警并提交 offset,避免阻塞后续消息
- DLT 失败等待时间默认限制为 1 秒
- 增加连续性放行、DLT 成功及失败监控指标
- 更新 Docker Compose、v3 升级文档及相关测试
2026-09-12 19:22:52 +08:00

118 lines
4.6 KiB
Bash
Executable File

#!/usr/bin/env bash
set -euo pipefail
# Docker 单文件版:直接在 Kafka 容器内创建并校验大日志相关 Topic。
# 默认值适配测试环境;如容器名、Kafka 地址或安装目录不同,可通过环境变量覆盖。
DOCKER_BIN="${DOCKER_BIN:-docker}"
KAFKA_CONTAINER="${KAFKA_CONTAINER:-kafka_ycgy-bt-kafka-1}"
BOOTSTRAP="${BOOTSTRAP:-localhost:9092}"
KAFKA_BIN="${KAFKA_BIN:-/opt/kafka/bin}"
RAW_TOPIC="${RAW_TOPIC:-agent-01-syslog-topic}"
LARGE_TOPIC="${LARGE_TOPIC:-${RAW_TOPIC}.large}"
MAX_MESSAGE_BYTES="${MAX_MESSAGE_BYTES:-33554432}"
RETENTION_MS="${RETENTION_MS:-604800000}"
DLT_RETENTION_MS="${DLT_RETENTION_MS:-2592000000}"
COMMAND_CONFIG="${COMMAND_CONFIG:-}"
die() {
echo "ERROR: $*" >&2
exit 1
}
command -v "${DOCKER_BIN}" >/dev/null 2>&1 || die "未找到 Docker CLI: ${DOCKER_BIN}"
RUNNING="$("${DOCKER_BIN}" inspect -f '{{.State.Running}}' "${KAFKA_CONTAINER}" 2>/dev/null || true)"
[[ "${RUNNING}" == "true" ]] || die "Kafka 容器不存在或未运行: ${KAFKA_CONTAINER}"
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" test -x "${KAFKA_BIN}/kafka-topics.sh" \
|| die "容器内找不到可执行文件: ${KAFKA_BIN}/kafka-topics.sh"
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" test -x "${KAFKA_BIN}/kafka-configs.sh" \
|| die "容器内找不到可执行文件: ${KAFKA_BIN}/kafka-configs.sh"
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" bash -c 'exit 0' \
|| die "容器内缺少 bash,无法执行原始升级脚本"
DOCKER_ENV=(
-e "BOOTSTRAP=${BOOTSTRAP}"
-e "KAFKA_BIN=${KAFKA_BIN}"
-e "RAW_TOPIC=${RAW_TOPIC}"
-e "LARGE_TOPIC=${LARGE_TOPIC}"
-e "MAX_MESSAGE_BYTES=${MAX_MESSAGE_BYTES}"
-e "RETENTION_MS=${RETENTION_MS}"
-e "DLT_RETENTION_MS=${DLT_RETENTION_MS}"
)
KAFKA_COMMON=(--bootstrap-server "${BOOTSTRAP}")
if [[ -n "${COMMAND_CONFIG}" ]]; then
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" test -r "${COMMAND_CONFIG}" \
|| die "容器内无法读取 COMMAND_CONFIG: ${COMMAND_CONFIG}"
DOCKER_ENV+=(-e "COMMAND_CONFIG=${COMMAND_CONFIG}")
KAFKA_COMMON+=(--command-config "${COMMAND_CONFIG}")
fi
echo "检查 Kafka 连接及原始 topic ..."
"${DOCKER_BIN}" exec "${KAFKA_CONTAINER}" \
"${KAFKA_BIN}/kafka-topics.sh" \
"${KAFKA_COMMON[@]}" \
--describe \
--topic "${RAW_TOPIC}" >/dev/null \
|| die "无法连接 Kafka 或原始 topic 不存在: ${RAW_TOPIC}"
echo "Kafka 容器: ${KAFKA_CONTAINER}"
echo "Bootstrap: ${BOOTSTRAP}"
echo "原始 topic: ${RAW_TOPIC}"
echo "大日志 topic: ${LARGE_TOPIC}"
echo "开始执行 Docker 版升级 ..."
exec "${DOCKER_BIN}" exec -i "${DOCKER_ENV[@]}" "${KAFKA_CONTAINER}" \
bash -s <<'KAFKA_UPGRADE_SCRIPT'
set -euo pipefail
TOPICS="${KAFKA_BIN}/kafka-topics.sh"
CONFIGS="${KAFKA_BIN}/kafka-configs.sh"
COMMON=(--bootstrap-server "${BOOTSTRAP}")
if [[ -n "${COMMAND_CONFIG:-}" ]]; then
[[ -r "${COMMAND_CONFIG}" ]] || { echo "ERROR: 无法读取 ${COMMAND_CONFIG}" >&2; exit 1; }
COMMON+=(--command-config "${COMMAND_CONFIG}")
fi
RAW_INFO=$("${TOPICS}" "${COMMON[@]}" --describe --topic "${RAW_TOPIC}")
PARTITIONS=$(printf '%s\n' "${RAW_INFO}" | awk '
/PartitionCount:/ {for(i=1;i<=NF;i++) if($i=="PartitionCount:"){print $(i+1);exit}}
')
REPLICAS=$(printf '%s\n' "${RAW_INFO}" | awk '
/ReplicationFactor:/ {for(i=1;i<=NF;i++) if($i=="ReplicationFactor:"){print $(i+1);exit}}
')
if [[ ! "${PARTITIONS}" =~ ^[1-9][0-9]*$ ]] || [[ ! "${REPLICAS}" =~ ^[1-9][0-9]*$ ]]; then
echo "ERROR: 无法读取 ${RAW_TOPIC} 的分区数或副本数" >&2
exit 1
fi
declare -a TOPIC_NAMES=(
"${LARGE_TOPIC}"
"${LARGE_TOPIC}.RETRY.1M"
"${LARGE_TOPIC}.RETRY.10M"
"${LARGE_TOPIC}.RETRY.1H"
"${LARGE_TOPIC}.DLT"
"${LARGE_TOPIC}.REPLAY"
)
echo "Kafka=${BOOTSTRAP},分区=${PARTITIONS},副本=${REPLICAS},单条上限=${MAX_MESSAGE_BYTES}"
for topic in "${TOPIC_NAMES[@]}"; do
topic_retention="${RETENTION_MS}"
[[ "${topic}" == "${LARGE_TOPIC}.DLT" ]] && topic_retention="${DLT_RETENTION_MS}"
"${TOPICS}" "${COMMON[@]}" --create --if-not-exists \
--topic "${topic}" --partitions "${PARTITIONS}" --replication-factor "${REPLICAS}" \
--config cleanup.policy=delete \
--config "retention.ms=${topic_retention}" \
--config "max.message.bytes=${MAX_MESSAGE_BYTES}"
"${CONFIGS}" "${COMMON[@]}" --alter --entity-type topics --entity-name "${topic}" \
--add-config "cleanup.policy=delete,retention.ms=${topic_retention},max.message.bytes=${MAX_MESSAGE_BYTES}"
"${TOPICS}" "${COMMON[@]}" --describe --topic "${topic}"
done
echo "PASS: 大日志 Topic 已创建并统一配置。"
echo "注意:Broker 的 replica.fetch.max.bytes 也必须不小于 ${MAX_MESSAGE_BYTES}。"
echo "脚本未删除消息,也未修改任何消费组 offset。"
KAFKA_UPGRADE_SCRIPT