fix(logging): 统一日志文件名校验规范-yuanrong-datasystem-AtomGit

M.repo_context/modules/infra/logging/design.md+6-3

M.repo_context/modules/infra/logging/log-lifecycle-and-rotation.md+3-0

	@@ -52,6 +52,9 @@
52	- `Logging::Start()` initializes client-specific config from environment when the explicit process role is	52	- `Logging::Start()` initializes client-specific config from environment when the explicit process role is
53	`LogProcessRole::CLIENT`; worker and coordinator callers pass `LogProcessRole::WORKER` and	53	`LogProcessRole::CLIENT`; worker and coordinator callers pass `LogProcessRole::WORKER` and
54	`LogProcessRole::COORDINATOR` respectively.	54	`LogProcessRole::COORDINATOR` respectively.
		55	+ - client log-name configuration keeps explicit `KVClientConfig` presence separate from the resolved `log_filename`;
		56	+ an explicit empty `LogName` or `AccessLogName` suppresses the corresponding environment override and selects the
		57	+ default client log base name.
55	- it sets `log_filename`, initializes the provider, starts `LogManager`, initializes `AccessRecorderManager`, and starts	58	- it sets `log_filename`, initializes the provider, starts `LogManager`, initializes `AccessRecorderManager`, and starts
56	the operation logger with an explicit `client`, `worker`, or `coordinator` role;	59	the operation logger with an explicit `client`, `worker`, or `coordinator` role;
57	- when `log_monitor` is enabled, `AccessRecorderManager` initializes exporters according to its construction-time	60	- when `log_monitor` is enabled, `AccessRecorderManager` initializes exporters according to its construction-time

Mcli/deploy/conf/coordinator_config.json+1-1

	@@ -49,7 +49,7 @@
49	},	49	},
50	"log_filename": {	50	"log_filename": {
51	"value": "datasystem_coordinator",	51	"value": "datasystem_coordinator",
52	- "description": "Prefix of log filename, default is program invocation short name. Use standard characters only."	52	+ "description": "Prefix of log filename. Only letters, digits, and underscores are allowed when non-empty."
53	},	53	},
54	"minloglevel": {	54	"minloglevel": {
55	"value": "0",	55	"value": "0",

Mcli/deploy/conf/worker_config.json+1-1

	@@ -441,7 +441,7 @@
441	},	441	},
442	"log_filename": {	442	"log_filename": {
443	"value": "",	443	"value": "",
444	- "description": "Prefix of log filename, default is program invocation short name. Use standard characters only."	444	+ "description": "Prefix of log filename. Only letters, digits, and underscores are allowed when non-empty."
445	},	445	},
446	"log_async_queue_size": {	446	"log_async_queue_size": {
447	"value": "65536",	447	"value": "65536",

Mdocs/source_zh_cn/api_reference/api/cpp/KVClientConfig.rst+6-3

Mdocs/source_zh_cn/appendix/client_env_guide.md+3-1

	@@ -39,7 +39,9 @@
39	\| 19 \| `DATASYSTEM_LOG_COMPRESS` \| `false` \| 日志文件是否压缩，默认不压缩。 \|	39	\| 19 \| `DATASYSTEM_LOG_COMPRESS` \| `false` \| 日志文件是否压缩，默认不压缩。 \|
40	\| 23 \| `DATASYSTEM_MIN_LOG_LEVEL` \| `0` \| 数据系统SDK的最小日志级别。数值越大，输出的日志越少；默认值为0，通常表示INFO级别及以上日志都可输出。 \|	40	\| 23 \| `DATASYSTEM_MIN_LOG_LEVEL` \| `0` \| 数据系统SDK的最小日志级别。数值越大，输出的日志越少；默认值为0，通常表示INFO级别及以上日志都可输出。 \|
41	\| 26 \| `DATASYSTEM_LOG_ONLY_WRITE_INFO_FILE` \| `true` \| INFO日志文件始终写入所有级别日志。该值为`true`时不额外生成WARNING/ERROR日志文件；为`false`时会额外生成WARNING/ERROR日志文件，高级别日志会按等级写入多个日志文件。 \|	41	\| 26 \| `DATASYSTEM_LOG_ONLY_WRITE_INFO_FILE` \| `true` \| INFO日志文件始终写入所有级别日志。该值为`true`时不额外生成WARNING/ERROR日志文件；为`false`时会额外生成WARNING/ERROR日志文件，高级别日志会按等级写入多个日志文件。 \|
42	-\| 28 \| `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` \| `true` \| 是否让客户端日志文件名不带进程号。默认`true`时输出为`ds_client.log`、`ds_client_access.log`；设置为`false`时恢复为带`<pid>`后缀的命名，适合多client进程同时运行且不希望覆盖日志的场景。 \|	42	+\| 28 \| `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` \| `false` \| 是否让客户端日志文件名不带进程号。默认`false`时输出带`<pid>`后缀的日志名；设置为`true`时输出为`ds_client.log`、`ds_client_access.log`。 \|
		43	+\| 33 \| `DATASYSTEM_CLIENT_LOG_NAME` \| `""` \| 客户端运行日志基名，非空时仅允许英文字母、数字和下划线。仅当 `KVClientConfig::LogName` 未显式配置时读取。 \|
		44	+\| 34 \| `DATASYSTEM_CLIENT_ACCESS_LOG_NAME` \| `""` \| 客户端访问日志基名，非空时仅允许英文字母、数字和下划线。仅当 `KVClientConfig::AccessLogName` 未显式配置时读取。 \|
43		45
44	## 运行时环境配置	46	## 运行时环境配置
45		47

Mdocs/source_zh_cn/appendix/log_guide.md+3-3

	@@ -29,9 +29,9 @@ openYuanrong datasystem 的日志分为以下类型：
29	\| 7 \| datasystem_worker \| `/path/yr_datasystem/logs/{log_filename}_operation.log` \| 操作审计日志；记录 Init/Shutdown、构建 Commit、配置初始化快照及运行时 flag 变更（`UpdateConfig` 或配置文件热更新） \|	29	\| 7 \| datasystem_worker \| `/path/yr_datasystem/logs/{log_filename}_operation.log` \| 操作审计日志；记录 Init/Shutdown、构建 Commit、配置初始化快照及运行时 flag 变更（`UpdateConfig` 或配置文件热更新） \|
30	\| 8 \| datasystem_worker \| `/path/yr_datasystem/logs/kv_resource.log` \| 资源快照 JSON-Lines；`resource.log` 同源字段的纯 JSON 输出，顶层 `time`/`pod_name`/`cluster_name`，`metrics` 按 schema 输出已配置的 metric 组；由 `json_log_monitor` 且 `log_monitor_exporter=harddisk` 控制 \|	30	\| 8 \| datasystem_worker \| `/path/yr_datasystem/logs/kv_resource.log` \| 资源快照 JSON-Lines；`resource.log` 同源字段的纯 JSON 输出，顶层 `time`/`pod_name`/`cluster_name`，`metrics` 按 schema 输出已配置的 metric 组；由 `json_log_monitor` 且 `log_monitor_exporter=harddisk` 控制 \|
31	\| 9 \| datasystem_worker \| `/path/yr_datasystem/logs/kv_metrics.log` \| 指标摘要 JSON-Lines；INFO.log 中 `metrics_summary` 的并行输出，body 逐字一致，额外前置 `time`/`pod_name`/`cluster_name`；由 `json_log_monitor` 控制（INFO.log 原输出仍由 `log_monitor` 控制） \|	31	\| 9 \| datasystem_worker \| `/path/yr_datasystem/logs/kv_metrics.log` \| 指标摘要 JSON-Lines；INFO.log 中 `metrics_summary` 的并行输出，body 逐字一致，额外前置 `time`/`pod_name`/`cluster_name`；由 `json_log_monitor` 控制（INFO.log 原输出仍由 `log_monitor` 控制） \|
32	-\| 10 \| Client \| `/path/client/ds_client.INFO.log`（及 `.WARNING`、`.ERROR` 等；关闭 `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` 后恢复为 `/path/client/ds_client_{pid}.INFO.log`） \| SDK 运行日志；基名可由启动参数与环境变量 `DATASYSTEM_CLIENT_LOG_NAME` 覆盖 \|	32	+\| 10 \| Client \| `/path/client/ds_client_{pid}.INFO.log`（及 `.WARNING`、`.ERROR` 等；开启 `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` 后为 `/path/client/ds_client.INFO.log`） \| SDK 运行日志；基名可由 `KVClientConfig::LogName` 与环境变量 `DATASYSTEM_CLIENT_LOG_NAME` 覆盖，非空时仅允许英文字母、数字和下划线；显式配置空串时忽略环境变量并使用默认基名 \|
33	-\| 11 \| Client \| `/path/client/ds_client_operation.log`（关闭 `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` 后恢复为 `/path/client/ds_client_{pid}_operation.log`） \| Client 操作审计日志；记录 Init/Shutdown 及 `UpdateConfig` 动态配置变更 \|	33	+\| 11 \| Client \| `/path/client/ds_client_{pid}_operation.log`（开启 `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` 后为 `/path/client/ds_client_operation.log`） \| Client 操作审计日志；记录 Init/Shutdown 及 `UpdateConfig` 动态配置变更 \|
34	-\| 12 \| Client \| `/path/client/ds_client_access.log`（关闭 `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` 后恢复为 `/path/client/ds_client_access_{pid}.log`） \| SDK 接口访问日志；基名可由 `DATASYSTEM_CLIENT_ACCESS_LOG_NAME` 覆盖 \|	34	+\| 12 \| Client \| `/path/client/ds_client_access_{pid}.log`（开启 `DATASYSTEM_CLIENT_LOG_WITHOUT_PID` 后为 `/path/client/ds_client_access.log`） \| SDK 接口访问日志；基名可由 `KVClientConfig::AccessLogName` 与环境变量 `DATASYSTEM_CLIENT_ACCESS_LOG_NAME` 覆盖，非空时仅允许英文字母、数字和下划线；显式配置空串时忽略环境变量并使用默认基名 \|
35	\| 13 \| datasystem_coordinator \| `/path/yr_datasystem/logs/datasystem_coordinator.INFO.log`（及 `.WARNING`、`.ERROR` 等轮转文件） \| Coordinator 运行日志；启动时记录 Git Commit 和 Branch \|	35	\| 13 \| datasystem_coordinator \| `/path/yr_datasystem/logs/datasystem_coordinator.INFO.log`（及 `.WARNING`、`.ERROR` 等轮转文件） \| Coordinator 运行日志；启动时记录 Git Commit 和 Branch \|
36	\| 14 \| datasystem_coordinator \| `/path/yr_datasystem/logs/datasystem_coordinator_operation.log` \| Coordinator 操作审计日志；记录 Init/Shutdown、构建 Commit、配置初始化及运行时 `UpdateConfig` 变更 \|	36	\| 14 \| datasystem_coordinator \| `/path/yr_datasystem/logs/datasystem_coordinator_operation.log` \| Coordinator 操作审计日志；记录 Init/Shutdown、构建 Commit、配置初始化及运行时 `UpdateConfig` 变更 \|
37		37

Mdocs/source_zh_cn/deployment/dscli.md+2-2

Minclude/datasystem/utils/kv_client_config.h+5-2

	@@ -62,7 +62,8 @@ public:
62		62
63	/**	63	/**
64	* @brief Client log file base name.	64	* @brief Client log file base name.
65	- * @param[in] name Equivalent to flag log_filename.	65	+ * @param[in] name Equivalent to flag log_filename. A non-empty name must contain only letters, digits, and
		66	+ * underscores. An empty name ignores DATASYSTEM_CLIENT_LOG_NAME and selects the default client log name.
66	* @return Reference to self for chaining.	67	* @return Reference to self for chaining.
67	*/	68	*/
68	Builder &LogName(const std::string &name);	69	Builder &LogName(const std::string &name);
	@@ -76,7 +77,9 @@ public:
76		77
77	/**	78	/**
78	* @brief Client access log file base name.	79	* @brief Client access log file base name.
79	- * @param[in] name Equivalent to client access log filename config.	80	+ * @param[in] name Equivalent to client access log filename config. A non-empty name must contain only letters,
		81	+ * digits, and underscores. An empty name ignores DATASYSTEM_CLIENT_ACCESS_LOG_NAME and selects the default
		82	+ * client access log name.
80	* @return Reference to self for chaining.	83	* @return Reference to self for chaining.
81	*/	84	*/
82	Builder &AccessLogName(const std::string &name);	85	Builder &AccessLogName(const std::string &name);

Mk8s/helm_chart/datasystem/values.yaml+1-1

	@@ -188,7 +188,7 @@ global:
188	# When enabled, requests whose RPC sub-phases exceed this threshold will include	188	# When enabled, requests whose RPC sub-phases exceed this threshold will include
189	# a latencySummary in the access log.	189	# a latencySummary in the access log.
190	slowLogRpcSlowerThan: 5000	190	slowLogRpcSlowerThan: 5000
191	- # Prefix of log filename, default is program invocation short name. Use standard characters only.	191	+ # Prefix of log filename. Only letters, digits, and underscores are allowed when non-empty.
192	logFilename: ""	192	logFilename: ""
193	# Interval between logging stream metrics	193	# Interval between logging stream metrics
194	scMetricsLogIntervalS: 60	194	scMetricsLogIntervalS: 60

Msrc/datasystem/client/object_cache/object_client_impl.cpp+12-6

	@@ -184,7 +184,8 @@ constexpr int64_t DRAINING_LOCATION_REFRESH_INITIAL_BACKOFF_MS = 1;
184	constexpr int32_t HASH_RING_RPC_MIN_TIMEOUT_MS = 100;	184	constexpr int32_t HASH_RING_RPC_MIN_TIMEOUT_MS = 100;
185	constexpr int BOUND_WORKER_PROBE_TIMEOUT_MS = 10;	185	constexpr int BOUND_WORKER_PROBE_TIMEOUT_MS = 10;
186	constexpr int TRANSPORT_DIAG_LOG_RATE = 100;	186	constexpr int TRANSPORT_DIAG_LOG_RATE = 100;
187	-const std::unordered_set<std::string> NON_GFLAG_KV_CLIENT_CONFIG_KEYS = {	187	+const std::unordered_set<std::string> KV_CLIENT_LOG_CONFIG_KEYS = {
		188	+ "log_filename",
188	"client_access_log_filename",	189	"client_access_log_filename",
189	"client_log_without_pid",	190	"client_log_without_pid",
190	};	191	};
	@@ -396,7 +397,7 @@ std::unordered_map<std::string, std::string> GetGflagArgs(const KVClientConfig &
396	{	397	{
397	std::unordered_map<std::string, std::string> args;	398	std::unordered_map<std::string, std::string> args;
398	for (const auto &arg : clientConfig.GetArgs()) {	399	for (const auto &arg : clientConfig.GetArgs()) {
399	- if (NON_GFLAG_KV_CLIENT_CONFIG_KEYS.find(arg.first) == NON_GFLAG_KV_CLIENT_CONFIG_KEYS.end()) {	400	+ if (KV_CLIENT_LOG_CONFIG_KEYS.find(arg.first) == KV_CLIENT_LOG_CONFIG_KEYS.end()) {
400	args.emplace(arg.first, arg.second);	401	args.emplace(arg.first, arg.second);
401	}	402	}
402	}	403	}
	@@ -405,12 +406,17 @@ std::unordered_map<std::string, std::string> GetGflagArgs(const KVClientConfig &
405		406
406	void ApplyKvClientLogConfig(const KVClientConfig &clientConfig)	407	void ApplyKvClientLogConfig(const KVClientConfig &clientConfig)
407	{	408	{
408	- auto logWithoutPid = clientConfig.GetArgs().find("client_log_without_pid");	409	+ const auto &args = clientConfig.GetArgs();
409	- if (logWithoutPid != clientConfig.GetArgs().end()) {	410	+ auto logName = args.find("log_filename");
		411	+ if (logName != args.end()) {
		412	+ Logging::SetClientLogName(logName->second);
		413	+ }
		414	+ auto logWithoutPid = args.find("client_log_without_pid");
		415	+ if (logWithoutPid != args.end()) {
410	Logging::SetClientLogWithoutPid(ParseBoolFromString(logWithoutPid->second, false));	416	Logging::SetClientLogWithoutPid(ParseBoolFromString(logWithoutPid->second, false));
411	}	417	}
412	- auto accessLogName = clientConfig.GetArgs().find("client_access_log_filename");	418	+ auto accessLogName = args.find("client_access_log_filename");
413	- if (accessLogName != clientConfig.GetArgs().end()) {	419	+ if (accessLogName != args.end()) {
414	Logging::SetClientAccessLogName(accessLogName->second);	420	Logging::SetClientAccessLogName(accessLogName->second);
415	}	421	}
416	}	422	}

Msrc/datasystem/common/flags/BUILD.bazel+0-1

	@@ -108,7 +108,6 @@ ds_cc_library(
108	"//include/datasystem/utils:utils_headers",	108	"//include/datasystem/utils:utils_headers",
109	"//src/datasystem/common/util:status_helper",	109	"//src/datasystem/common/util:status_helper",
110	"//src/datasystem/common/util:validator",	110	"//src/datasystem/common/util:validator",
111	- "@re2",
112	],	111	],
113	alwayslink = True,	112	alwayslink = True,
114	)	113	)

Msrc/datasystem/common/flags/common_flag_define.cpp+2-1

	@@ -40,7 +40,8 @@ DS_DEFINE_string(
40	unix_domain_socket_dir, "~/datasystem/unix_domain_socket_dir",	40	unix_domain_socket_dir, "~/datasystem/unix_domain_socket_dir",
41	"The directory to store unix domain socket file. The UDS generates temporary files in this path. Max lenth: 80");	41	"The directory to store unix domain socket file. The UDS generates temporary files in this path. Max lenth: 80");
42	DS_DEFINE_string(log_filename, "",	42	DS_DEFINE_string(log_filename, "",
43	- "Prefix of log filename, default is program invocation short name. Use standard characters only.");	43	+ "Prefix of log filename, default is program invocation short name. Only letters, digits, and "
		44	+ "underscores are allowed.");
44	DS_DEFINE_string(curve_key_dir, "",	45	DS_DEFINE_string(curve_key_dir, "",
45	"The directory to find ZMQ curve key files. This path must be specified "	46	"The directory to find ZMQ curve key files. This path must be specified "
46	"when zmq authentication is enabled. Path must be less than 4095 characters (PATH_MAX).");	47	"when zmq authentication is enabled. Path must be less than 4095 characters (PATH_MAX).");

Msrc/datasystem/common/flags/common_flags_validate.cpp+11-1

	@@ -222,6 +222,16 @@ bool ValidateSampleRateRange(const char *flagName, double value)
222	}	222	}
223	return true;	223	return true;
224	}	224	}
		225	+
		226	+bool ValidateOptionalLogName(const char *flagName, const std::string &value)
		227	+{
		228	+ if (value.empty() \|\| Validator::ValidateLogName(value)) {
		229	+ return true;
		230	+ }
		231	+ LOG(ERROR) << FormatString("The value of %s flag is %s, which must contain only a-z, A-Z, 0-9, and underscore.",
		232	+ flagName, value);
		233	+ return false;
		234	+}
225	} // namespace	235	} // namespace
226		236
227	DS_DEFINE_validator(l2_cache_type, &Validator::ValidateL2CacheType);	237	DS_DEFINE_validator(l2_cache_type, &Validator::ValidateL2CacheType);
	@@ -253,7 +263,7 @@ DS_DEFINE_validator(rebalance_usage_gap_percent, &ValidatePercent);
253	DS_DEFINE_validator(rebalance_task_report_grace_ms, &Validator::ValidateUint32);	263	DS_DEFINE_validator(rebalance_task_report_grace_ms, &Validator::ValidateUint32);
254	DS_DEFINE_validator(monitor_config_file, &Validator::ValidatePathString);	264	DS_DEFINE_validator(monitor_config_file, &Validator::ValidatePathString);
255	DS_DEFINE_validator(unix_domain_socket_dir, &Validator::ValidateUnixDomainSocketDir);	265	DS_DEFINE_validator(unix_domain_socket_dir, &Validator::ValidateUnixDomainSocketDir);
256	-DS_DEFINE_validator(log_filename, &Validator::ValidateEligibleChar);	266	+DS_DEFINE_validator(log_filename, &ValidateOptionalLogName);
257	DS_DEFINE_validator(curve_key_dir, &Validator::ValidatePathString);	267	DS_DEFINE_validator(curve_key_dir, &Validator::ValidatePathString);
258	DS_DEFINE_validator(shared_disk_directory, &Validator::ValidatePathString);	268	DS_DEFINE_validator(shared_disk_directory, &Validator::ValidatePathString);
259	DS_DEFINE_validator(distributed_disk_path, &Validator::ValidatePathString);	269	DS_DEFINE_validator(distributed_disk_path, &Validator::ValidatePathString);

Msrc/datasystem/common/flags/kv_client_config.cpp+7-19

	@@ -18,8 +18,6 @@
18	#include <sstream>	18	#include <sstream>
19	#include <vector>	19	#include <vector>
20		20
21	-#include "re2/re2.h"
22	-
23	#include "datasystem/common/util/validator.h"	21	#include "datasystem/common/util/validator.h"
24		22
25	namespace datasystem {	23	namespace datasystem {
	@@ -95,21 +93,15 @@ bool ValidateNonEmptyArg(const std::unordered_map<std::string, std::string> &arg
95	return it == args.end() \|\| !it->second.empty();	93	return it == args.end() \|\| !it->second.empty();
96	}	94	}
97		95
98	-bool IsValidNonEmptyLogBaseName(const std::string &value)	96	+void ValidateLogNameArg(const std::unordered_map<std::string, std::string> &args, const std::string &key,
		97	+ const std::string &fieldName, std::vector<std::string> &errors)
99	{	98	{
100	- static const re2::RE2 pattern("^[a-zA-Z0-9_]*$");	99	+ auto it = args.find(key);
101	- return re2::RE2::FullMatch(value, pattern);
102	-}
103	-
104	-void ValidateAccessLogNameArg(const std::unordered_map<std::string, std::string> &args,
105	- std::vector<std::string> &errors)
106	-{
107	- auto it = args.find("client_access_log_filename");
108	if (it == args.end() \|\| it->second.empty()) {	100	if (it == args.end() \|\| it->second.empty()) {
109	return;	101	return;
110	}	102	}
111	- if (!IsValidNonEmptyLogBaseName(it->second)) {	103	+ if (!Validator::ValidateLogName(it->second)) {
112	- AddError(errors, "AccessLogName", "must contain only a-z, A-Z, 0-9, and underscore");	104	+ AddError(errors, fieldName, "must contain only a-z, A-Z, 0-9, and underscore");
113	}	105	}
114	}	106	}
115		107
	@@ -140,12 +132,8 @@ void ValidateKvClientStringFields(const std::unordered_map<std::string, std::str
140	std::vector<std::string> &errors)	132	std::vector<std::string> &errors)
141	{	133	{
142	ValidateNonEmptyPathArg(args, "log_dir", "LogDir", errors);	134	ValidateNonEmptyPathArg(args, "log_dir", "LogDir", errors);
143	- if (!ValidateNonEmptyArg(args, "log_filename")) {	135	+ ValidateLogNameArg(args, "log_filename", "LogName", errors);
144	- AddError(errors, "LogName", "must not be empty");	136	+ ValidateLogNameArg(args, "client_access_log_filename", "AccessLogName", errors);
145	- } else if (!ValidateStringArg(args, "log_filename", &Validator::ValidateEligibleChar)) {
146	- AddError(errors, "LogName", "contains unsupported characters");
147	- }
148	- ValidateAccessLogNameArg(args, errors);
149	ValidatePathArg(args, "monitor_config_file", "MonitorConfigPath", errors, true);	137	ValidatePathArg(args, "monitor_config_file", "MonitorConfigPath", errors, true);
150	}	138	}
151		139

Msrc/datasystem/common/log/BUILD.bazel+0-1

	@@ -142,7 +142,6 @@ ds_cc_library(
142	"//src/datasystem/common/util:request_context",	142	"//src/datasystem/common/util:request_context",
143	"//src/datasystem/common/util:thread",	143	"//src/datasystem/common/util:thread",
144	"//src/datasystem/common/util:uuid_generator",	144	"//src/datasystem/common/util:uuid_generator",
145	- "@re2",
146	"@securec",	145	"@securec",
147	],	146	],
148	alwayslink = 1,	147	alwayslink = 1,

Msrc/datasystem/common/log/access_recorder.cpp+2-1

	@@ -38,6 +38,7 @@
38	#include "datasystem/common/util/format.h"	38	#include "datasystem/common/util/format.h"
39	#include "datasystem/common/util/request_context.h"	39	#include "datasystem/common/util/request_context.h"
40	#include "datasystem/common/util/status_helper.h"	40	#include "datasystem/common/util/status_helper.h"
		41	+#include "datasystem/common/util/validator.h"
41	#include "datasystem/utils/status.h"	42	#include "datasystem/utils/status.h"
42		43
43	DS_DECLARE_bool(log_monitor);	44	DS_DECLARE_bool(log_monitor);
	@@ -989,7 +990,7 @@ Status AccessRecorderManager::ResetWriteLogger(bool isEmbeddedClient)
989		990
990	// Allow overriding client access log filename via config or environment variable.	991	// Allow overriding client access log filename via config or environment variable.
991	std::string accessLogName = Logging::GetClientAccessLogName();	992	std::string accessLogName = Logging::GetClientAccessLogName();
992	- if (Logging::ValidateLogName(accessLogName)) {	993	+ if (Validator::ValidateLogName(accessLogName)) {
993	clientAccessLogName = std::move(accessLogName);	994	clientAccessLogName = std::move(accessLogName);
994	}	995	}
995		996

Msrc/datasystem/common/log/logging.cpp+26-13

	@@ -28,7 +28,6 @@
28	#include <fcntl.h>	28	#include <fcntl.h>
29	#include <unistd.h>	29	#include <unistd.h>
30		30
31	-#include "re2/re2.h"
32	#include "datasystem/common/log/access_recorder.h"	31	#include "datasystem/common/log/access_recorder.h"
33	#include "datasystem/common/metrics/hard_disk_exporter/hard_disk_exporter.h"	32	#include "datasystem/common/metrics/hard_disk_exporter/hard_disk_exporter.h"
34	#include "datasystem/common/log/log.h"	33	#include "datasystem/common/log/log.h"
	@@ -141,6 +140,8 @@ namespace {
141	std::mutex g_clientLogConfigMutex;	140	std::mutex g_clientLogConfigMutex;
142	bool g_hasClientLogWithoutPidConfig = false;	141	bool g_hasClientLogWithoutPidConfig = false;
143	bool g_clientLogWithoutPidConfig = false;	142	bool g_clientLogWithoutPidConfig = false;
		143	+bool g_hasClientLogNameConfig = false;
		144	+std::string g_clientLogNameConfig;
144	bool g_hasClientAccessLogNameConfig = false;	145	bool g_hasClientAccessLogNameConfig = false;
145	std::string g_clientAccessLogNameConfig;	146	std::string g_clientAccessLogNameConfig;
146	} // namespace	147	} // namespace
	@@ -481,6 +482,23 @@ void Logging::SetClientLogWithoutPid(bool enabled)
481	g_clientLogWithoutPidConfig = enabled;	482	g_clientLogWithoutPidConfig = enabled;
482	}	483	}
483		484
		485	+void Logging::SetClientLogName(const std::string &logName)
		486	+{
		487	+ std::lock_guard<std::mutex> lock(g_clientLogConfigMutex);
		488	+ g_hasClientLogNameConfig = true;
		489	+ g_clientLogNameConfig = logName;
		490	+}
		491	+
		492	+bool Logging::TryGetClientLogName(std::string &logName)
		493	+{
		494	+ std::lock_guard<std::mutex> lock(g_clientLogConfigMutex);
		495	+ if (!g_hasClientLogNameConfig) {
		496	+ return false;
		497	+ }
		498	+ logName = g_clientLogNameConfig;
		499	+ return true;
		500	+}
		501	+
484	void Logging::SetClientAccessLogName(const std::string &logName)	502	void Logging::SetClientAccessLogName(const std::string &logName)
485	{	503	{
486	std::lock_guard<std::mutex> lock(g_clientLogConfigMutex);	504	std::lock_guard<std::mutex> lock(g_clientLogConfigMutex);
	@@ -503,6 +521,8 @@ void Logging::ResetClientLogConfigForTest()
503	std::lock_guard<std::mutex> lock(g_clientLogConfigMutex);	521	std::lock_guard<std::mutex> lock(g_clientLogConfigMutex);
504	g_hasClientLogWithoutPidConfig = false;	522	g_hasClientLogWithoutPidConfig = false;
505	g_clientLogWithoutPidConfig = false;	523	g_clientLogWithoutPidConfig = false;
		524	+ g_hasClientLogNameConfig = false;
		525	+ g_clientLogNameConfig.clear();
506	g_hasClientAccessLogNameConfig = false;	526	g_hasClientAccessLogNameConfig = false;
507	g_clientAccessLogNameConfig.clear();	527	g_clientAccessLogNameConfig.clear();
508	}	528	}
	@@ -552,14 +572,17 @@ void Logging::Start(const std::string &logFilename, LogProcessRole processRole,
552	if (isClient_) {	572	if (isClient_) {
553	GetInstance()->InitClientConfig();	573	GetInstance()->InitClientConfig();
554		574
555	- if (!FLAGS_log_filename.empty()) {	575	+ std::string configuredLogName;
		576	+ if (TryGetClientLogName(configuredLogName)) {
		577	+ clientLogName = configuredLogName.empty() ? GetClientLogName(logFilename, getpid()) : configuredLogName;
		578	+ } else if (!FLAGS_log_filename.empty()) {
556	clientLogName = FLAGS_log_filename;	579	clientLogName = FLAGS_log_filename;
557	} else {	580	} else {
558	clientLogName = GetClientLogName(logFilename, getpid());	581	clientLogName = GetClientLogName(logFilename, getpid());
559		582
560	// Allow overriding client log filename via environment variable	583	// Allow overriding client log filename via environment variable
561	std::string logName = GetStringFromEnv(LOG_NAME_ENV.c_str(), "");	584	std::string logName = GetStringFromEnv(LOG_NAME_ENV.c_str(), "");
562	- if (ValidateLogName(logName)) {	585	+ if (Validator::ValidateLogName(logName)) {
563	clientLogName = std::move(logName);	586	clientLogName = std::move(logName);
564	}	587	}
565	}	588	}
	@@ -630,14 +653,4 @@ Status Logging::WriteLogToFile(int lineOfCode, const std::string &fileNameOfCode
630	return Status::OK();	653	return Status::OK();
631	}	654	}
632		655
633	-bool Logging::ValidateLogName(const std::string &logName)
634	-{
635	- if (logName.empty()) {
636	- return false;
637	- }
638	-
639	- // Only allow: a-z, A-Z, 0-9, _
640	- static const re2::RE2 re("^[a-zA-Z0-9_]*$");
641	- return re2::RE2::FullMatch(logName, re);
642	-}
643	} // namespace datasystem	656	} // namespace datasystem

Msrc/datasystem/common/log/logging.h+13-2

	@@ -126,8 +126,6 @@ public:
126	return podName_;	126	return podName_;
127	}	127	}
128		128
129	- static bool ValidateLogName(const std::string &logName);
130	-
131	/**	129	/**
132	* @brief Build client log filename.	130	* @brief Build client log filename.
133	* @param[in] baseName Log file base name.	131	* @param[in] baseName Log file base name.
	@@ -148,6 +146,19 @@ public:
148	*/	146	*/
149	static void SetClientLogWithoutPid(bool enabled);	147	static void SetClientLogWithoutPid(bool enabled);
150		148
		149	+ /**
		150	+ * @brief Set client log filename configured by KVClientConfig.
		151	+ * @param[in] logName Client log file base name. An empty value selects the default name.
		152	+ */
		153	+ static void SetClientLogName(const std::string &logName);
		154	+
		155	+ /**
		156	+ * @brief Get client log filename explicitly configured by KVClientConfig.
		157	+ * @param[out] logName Client log file base name, which may be empty.
		158	+ * @return True if KVClientConfig explicitly configured the log name.
		159	+ */
		160	+ static bool TryGetClientLogName(std::string &logName);
		161	+
151	/**	162	/**
152	* @brief Set client access log filename.	163	* @brief Set client access log filename.
153	* @param[in] logName Access log file base name.	164	* @param[in] logName Access log file base name.

Msrc/datasystem/common/util/validator.h+11-0

	@@ -434,6 +434,17 @@ public:
434	return false;	434	return false;
435	}	435	}
436		436
		437	+ /**
		438	+ * @brief Validate a log file base name.
		439	+ * @param[in] value The string to be checked.
		440	+ * @return True if valid.
		441	+ */
		442	+ static bool ValidateLogName(const std::string &value)
		443	+ {
		444	+ static const re2::RE2 re("^[a-zA-Z0-9_]+$");
		445	+ return re2::RE2::FullMatch(value, re);
		446	+ }
		447	+
437	/**	448	/**
438	* @brief Validate the given string contains all legal characters.	449	* @brief Validate the given string contains all legal characters.
439	* @param[in] flagName The characters flag.	450	* @param[in] flagName The characters flag.

Mtests/perf/zmq/zmq_perf_agt_exe.cpp+1-1

	@@ -44,7 +44,7 @@ void SignalHandler(int signum)
44	int main(int argc, char **argv)	44	int main(int argc, char **argv)
45	{	45	{
46	auto pid = getpid();	46	auto pid = getpid();
47	- FLAGS_log_filename = datasystem::FormatString("%s.%d", "zmq_perf_agent", pid);	47	+ FLAGS_log_filename = datasystem::FormatString("%s_%d", "zmq_perf_agent", pid);
48	FLAGS_v = datasystem::RPC_KEY_LOG_LEVEL;	48	FLAGS_v = datasystem::RPC_KEY_LOG_LEVEL;
49	datasystem::Logging::GetInstance()->Start(FLAGS_log_filename, datasystem::LogProcessRole::WORKER);	49	datasystem::Logging::GetInstance()->Start(FLAGS_log_filename, datasystem::LogProcessRole::WORKER);
50	datasystem::st::ZmqPerfAgent zmqPerfAgent;	50	datasystem::st::ZmqPerfAgent zmqPerfAgent;

Mtests/st/client/kv_cache/kv_client_init_test.cpp+31-0

	@@ -26,6 +26,7 @@
26		26
27	#include "client/object_cache/oc_client_common.h"	27	#include "client/object_cache/oc_client_common.h"
28	#include "common.h"	28	#include "common.h"
		29	+#include "datasystem/common/constants.h"
29	#include "datasystem/common/flags/flags.h"	30	#include "datasystem/common/flags/flags.h"
30	#include "datasystem/common/inject/inject_point.h"	31	#include "datasystem/common/inject/inject_point.h"
31	#include "datasystem/common/flags/common_flags.h"	32	#include "datasystem/common/flags/common_flags.h"
	@@ -37,6 +38,7 @@ DS_DECLARE_uint32(max_log_size);
37	DS_DECLARE_uint32(log_async_queue_size);	38	DS_DECLARE_uint32(log_async_queue_size);
38	DS_DECLARE_int32(v);	39	DS_DECLARE_int32(v);
39	DS_DECLARE_string(monitor_config_file);	40	DS_DECLARE_string(monitor_config_file);
		41	+DS_DECLARE_string(log_filename);
40		42
41	namespace datasystem {	43	namespace datasystem {
42	namespace st {	44	namespace st {
	@@ -215,6 +217,35 @@ TEST_F(KVClientInitTest, ConfigAfterDefaultInitDoesNotOverrideProcessConfig)
215	});	217	});
216	}	218	}
217		219
		220	+TEST_F(KVClientInitTest, ExplicitEmptyLogNameUsesDefaultAndIgnoresEnv)
		221	+{
		222	+ if (FLAGS_use_brpc) {
		223	+ GTEST_SKIP() << "brpc fork-safety: brpc channel/bthread global state not fork-safe. Tracked separately.";
		224	+ }
		225	+ auto connectOptions = GetConnectOptions();
		226	+ RunInChildProcess([connectOptions]() -> int {
		227	+ setenv("DATASYSTEM_CLIENT_LOG_NAME", "env_client", 1);
		228	+ KVClientConfig config;
		229	+ auto buildStatus = KVClientConfig::Builder().LogName("").LogWithoutPid(true).Build(config);
		230	+ if (buildStatus.IsError()) {
		231	+ LOG(ERROR) << "Build KVClientConfig failed: " << buildStatus.ToString();
		232	+ return 1;
		233	+ }
		234	+ KVClient client(connectOptions);
		235	+ auto initStatus = client.Init(config);
		236	+ if (initStatus.IsError()) {
		237	+ LOG(ERROR) << "Init with explicit empty LogName failed: " << initStatus.ToString();
		238	+ return 1;
		239	+ }
		240	+ if (FLAGS_log_filename != CLIENT_LOG_FILENAME) {
		241	+ LOG(ERROR) << "Unexpected log_filename, expect: " << CLIENT_LOG_FILENAME
		242	+ << ", actual: " << FLAGS_log_filename;
		243	+ return 1;
		244	+ }
		245	+ return 0;
		246	+ });
		247	+}
		248	+
218	TEST_F(KVClientInitTest, EmptyMonitorConfigPathAllowsUpdateConfig)	249	TEST_F(KVClientInitTest, EmptyMonitorConfigPathAllowsUpdateConfig)
219	{	250	{
220	if (FLAGS_use_brpc) {	251	if (FLAGS_use_brpc) {

Mtests/ut/common/flags/flags_test.cpp+84-10

	@@ -60,6 +60,7 @@ DS_DECLARE_uint32(log_retention_day);
60	DS_DECLARE_uint32(max_log_file_num);	60	DS_DECLARE_uint32(max_log_file_num);
61	DS_DECLARE_uint32(max_log_size);	61	DS_DECLARE_uint32(max_log_size);
62	DS_DECLARE_uint32(stderrthreshold);	62	DS_DECLARE_uint32(stderrthreshold);
		63	+DS_DECLARE_string(log_filename);
63	DS_DECLARE_string(monitor_config_file);	64	DS_DECLARE_string(monitor_config_file);
64		65
65	namespace {	66	namespace {
	@@ -282,6 +283,10 @@ private:
282	void ApplyKvClientLogConfigFromConfig(const KVClientConfig &clientConfig)	283	void ApplyKvClientLogConfigFromConfig(const KVClientConfig &clientConfig)
283	{	284	{
284	const auto &args = clientConfig.GetArgs();	285	const auto &args = clientConfig.GetArgs();
		286	+ auto logName = args.find("log_filename");
		287	+ if (logName != args.end()) {
		288	+ Logging::SetClientLogName(logName->second);
		289	+ }
285	auto logWithoutPid = args.find("client_log_without_pid");	290	auto logWithoutPid = args.find("client_log_without_pid");
286	if (logWithoutPid != args.end()) {	291	if (logWithoutPid != args.end()) {
287	Logging::SetClientLogWithoutPid(ParseBoolFromString(logWithoutPid->second, false));	292	Logging::SetClientLogWithoutPid(ParseBoolFromString(logWithoutPid->second, false));
	@@ -298,6 +303,11 @@ protected:
298	{	303	{
299	Logging::ResetClientLogConfigForTest();	304	Logging::ResetClientLogConfigForTest();
300	}	305	}
		306	+
		307	+ void TearDown() override
		308	+ {
		309	+ Logging::ResetClientLogConfigForTest();
		310	+ }
301	};	311	};
302	} // namespace	312	} // namespace
303		313
	@@ -479,6 +489,27 @@ TEST_F(ClientLogConfigPriorityTest, ExplicitClientAccessLogNameIsNotOverriddenBy
479	}	489	}
480	}	490	}
481		491
		492	+TEST_F(ClientLogConfigPriorityTest, ExplicitEmptyClientLogNameIsPreserved)
		493	+{
		494	+ KVClientConfig config;
		495	+ ASSERT_EQ(KVClientConfig::Builder().LogName("").Build(config), Status::OK());
		496	+ ApplyKvClientLogConfigFromConfig(config);
		497	+
		498	+ std::string logName;
		499	+ ASSERT_TRUE(Logging::TryGetClientLogName(logName));
		500	+ ASSERT_TRUE(logName.empty());
		501	+}
		502	+
		503	+TEST_F(ClientLogConfigPriorityTest, UnspecifiedClientLogNameHasNoExplicitConfig)
		504	+{
		505	+ KVClientConfig config;
		506	+ ASSERT_EQ(KVClientConfig::Builder().Build(config), Status::OK());
		507	+ ApplyKvClientLogConfigFromConfig(config);
		508	+
		509	+ std::string logName;
		510	+ ASSERT_FALSE(Logging::TryGetClientLogName(logName));
		511	+}
		512	+
482	TEST_F(ClientLogConfigPriorityTest, ExplicitEmptyClientAccessLogNameIsNotOverriddenByEnv)	513	TEST_F(ClientLogConfigPriorityTest, ExplicitEmptyClientAccessLogNameIsNotOverriddenByEnv)
483	{	514	{
484	const char *envName = ACCESS_LOG_NAME_ENV.c_str();	515	const char *envName = ACCESS_LOG_NAME_ENV.c_str();
	@@ -630,16 +661,35 @@ TEST_F(FlagsTest, KVClientConfigBuilderAggregatesInvalidValues)
630	EXPECT_TRUE(config.GetArgs().empty());	661	EXPECT_TRUE(config.GetArgs().empty());
631	}	662	}
632		663
633	-TEST_F(FlagsTest, KVClientConfigBuilderRejectsEmptyGflagStrings)	664	+TEST_F(FlagsTest, KVClientConfigBuilderValidatesLogName)
		665	+{
		666	+ KVClientConfig config;
		667	+ Status status = KVClientConfig::Builder().LogName("client_log_01").Build(config);
		668	+ ASSERT_EQ(status, Status::OK());
		669	+ ASSERT_EQ(config.GetArgs().at("log_filename"), "client_log_01");
		670	+
		671	+ KVClientConfig emptyConfig;
		672	+ status = KVClientConfig::Builder().LogName("").Build(emptyConfig);
		673	+ ASSERT_EQ(status, Status::OK());
		674	+ ASSERT_EQ(emptyConfig.GetArgs().at("log_filename"), "");
		675	+
		676	+ const std::vector<std::string> invalidLogNames = { "client-log", "a b", "a/b", "../client",
		677	+ "client.log", "client:name", "客户端" };
		678	+ for (const auto &logName : invalidLogNames) {
		679	+ KVClientConfig invalidConfig;
		680	+ status = KVClientConfig::Builder().LogName(logName).Build(invalidConfig);
		681	+ EXPECT_EQ(status.GetCode(), StatusCode::K_INVALID) << logName;
		682	+ EXPECT_THAT(status.GetMsg(), testing::HasSubstr("LogName")) << logName;
		683	+ EXPECT_TRUE(invalidConfig.GetArgs().empty()) << logName;
		684	+ }
		685	+}
		686	+
		687	+TEST_F(FlagsTest, KVClientConfigBuilderRejectsEmptyRequiredGflagStrings)
634	{	688	{
635	KVClientConfig config;	689	KVClientConfig config;
636	Status status = KVClientConfig::Builder().LogDir("").Build(config);	690	Status status = KVClientConfig::Builder().LogDir("").Build(config);
637	ASSERT_EQ(status.GetCode(), StatusCode::K_INVALID);	691	ASSERT_EQ(status.GetCode(), StatusCode::K_INVALID);
638	EXPECT_THAT(status.GetMsg(), testing::HasSubstr("LogDir"));	692	EXPECT_THAT(status.GetMsg(), testing::HasSubstr("LogDir"));
639	-
640	- status = KVClientConfig::Builder().LogName("").Build(config);
641	- ASSERT_EQ(status.GetCode(), StatusCode::K_INVALID);
642	- EXPECT_THAT(status.GetMsg(), testing::HasSubstr("LogName"));
643	}	693	}
644		694
645	TEST_F(FlagsTest, KVClientConfigBuilderAcceptsEmptyMonitorConfigPath)	695	TEST_F(FlagsTest, KVClientConfigBuilderAcceptsEmptyMonitorConfigPath)
	@@ -682,13 +732,37 @@ TEST_F(FlagsTest, ParseCommandLineFlagsRejectsEmptyStringWhenValidatorDisallows)
682	EXPECT_THAT(errMsg, testing::HasSubstr("illegal value"));	732	EXPECT_THAT(errMsg, testing::HasSubstr("illegal value"));
683	}	733	}
684		734
685	-TEST_F(FlagsTest, KVClientConfigBuilderRejectsInvalidAccessLogName)	735	+TEST_F(FlagsTest, KVClientConfigBuilderValidatesAccessLogName)
686	{	736	{
687	KVClientConfig config;	737	KVClientConfig config;
688	- Status status = KVClientConfig::Builder().AccessLogName("client-access").Build(config);	738	+ Status status = KVClientConfig::Builder().AccessLogName("client_access_01").Build(config);
689	- ASSERT_EQ(status.GetCode(), StatusCode::K_INVALID);	739	+ ASSERT_EQ(status, Status::OK());
690	- EXPECT_THAT(status.GetMsg(), testing::HasSubstr("AccessLogName"));	740	+ ASSERT_EQ(config.GetArgs().at("client_access_log_filename"), "client_access_01");
691	- EXPECT_TRUE(config.GetArgs().empty());	741	+
		742	+ KVClientConfig emptyConfig;
		743	+ status = KVClientConfig::Builder().AccessLogName("").Build(emptyConfig);
		744	+ ASSERT_EQ(status, Status::OK());
		745	+ ASSERT_EQ(emptyConfig.GetArgs().at("client_access_log_filename"), "");
		746	+
		747	+ for (const auto &logName : { "client-access", "client/access", "client.access" }) {
		748	+ KVClientConfig invalidConfig;
		749	+ status = KVClientConfig::Builder().AccessLogName(logName).Build(invalidConfig);
		750	+ EXPECT_EQ(status.GetCode(), StatusCode::K_INVALID) << logName;
		751	+ EXPECT_THAT(status.GetMsg(), testing::HasSubstr("AccessLogName")) << logName;
		752	+ EXPECT_TRUE(invalidConfig.GetArgs().empty()) << logName;
		753	+ }
		754	+}
		755	+
		756	+TEST_F(FlagsTest, LogFilenameFlagUsesUnifiedCharacterContract)
		757	+{
		758	+ const std::string originalLogFilename = FLAGS_log_filename;
		759	+ std::string errMsg;
		760	+ EXPECT_TRUE(SetCommandLineOption("log_filename", "client_log_01", errMsg));
		761	+ EXPECT_FALSE(SetCommandLineOption("log_filename", "client-log", errMsg));
		762	+ EXPECT_EQ(FLAGS_log_filename, "client_log_01");
		763	+ EXPECT_TRUE(SetCommandLineOption("log_filename", "", errMsg));
		764	+ EXPECT_TRUE(FLAGS_log_filename.empty());
		765	+ FLAGS_log_filename = originalLogFilename;
692	}	766	}
693		767
694	TEST_F(FlagsTest, ParseCommandLineFlagsClearsErrorStateOnRetry)	768	TEST_F(FlagsTest, ParseCommandLineFlagsClearsErrorStateOnRetry)

Mtests/ut/common/log/BUILD.bazel+1-0

	@@ -36,6 +36,7 @@ ds_cc_test(
36	tags = ["manual"],	36	tags = ["manual"],
37	deps = [	37	deps = [
38	"//src/datasystem/common/log:common_log",	38	"//src/datasystem/common/log:common_log",
		39	+ "//src/datasystem/common/util:validator",
39	"//tests/ut:ut_common",	40	"//tests/ut:ut_common",
40	],	41	],
41	)	42	)

Mtests/ut/common/log/logging_test.cpp+34-1

	@@ -54,6 +54,7 @@
54	#include "datasystem/common/util/thread_pool.h"	54	#include "datasystem/common/util/thread_pool.h"
55	#include "datasystem/common/util/timer.h"	55	#include "datasystem/common/util/timer.h"
56	#include "datasystem/common/util/uri.h"	56	#include "datasystem/common/util/uri.h"
		57	+#include "datasystem/common/util/validator.h"
57	#include "datasystem/utils/status.h"	58	#include "datasystem/utils/status.h"
58		59
59	DS_DECLARE_bool(alsologtostderr);	60	DS_DECLARE_bool(alsologtostderr);
	@@ -157,7 +158,7 @@ public:
157	std::string ClientAccessLogName()	158	std::string ClientAccessLogName()
158	{	159	{
159	std::string accessLogName = GetStringFromEnv(ACCESS_LOG_NAME_ENV.c_str(), "");	160	std::string accessLogName = GetStringFromEnv(ACCESS_LOG_NAME_ENV.c_str(), "");
160	- if (Logging::ValidateLogName(accessLogName)) {	161	+ if (Validator::ValidateLogName(accessLogName)) {
161	return accessLogName;	162	return accessLogName;
162	}	163	}
163	return Logging::GetClientLogName(CLIENT_ACCESS_LOG_NAME, getpid());	164	return Logging::GetClientLogName(CLIENT_ACCESS_LOG_NAME, getpid());
	@@ -1057,5 +1058,37 @@ TEST_F(LoggingTest, TestLogName)
1057	filepath = FLAGS_log_dir + "/test_client.INFO.log";	1058	filepath = FLAGS_log_dir + "/test_client.INFO.log";
1058	ASSERT_TRUE(FileExist(filepath));	1059	ASSERT_TRUE(FileExist(filepath));
1059	}	1060	}
		1061	+
		1062	+TEST_F(LoggingTest, TestExplicitEmptyClientLogNameUsesDefault)
		1063	+{
		1064	+ ScopedEnv logName(LOG_NAME_ENV, "env_client");
		1065	+ ScopedEnv clientLogWithoutPid(CLIENT_LOG_WITHOUT_PID_ENV, "true");
		1066	+ Logging::SetClientLogName("");
		1067	+
		1068	+ Logging::GetInstance()->Start(CLIENT_LOG_FILENAME, LogProcessRole::CLIENT, 1);
		1069	+
		1070	+ ASSERT_EQ(FLAGS_log_filename, CLIENT_LOG_FILENAME);
		1071	+}
		1072	+
		1073	+TEST_F(LoggingTest, TestExplicitClientLogNameOverridesEnv)
		1074	+{
		1075	+ ScopedEnv logName(LOG_NAME_ENV, "env_client");
		1076	+ Logging::SetClientLogName("config_client");
		1077	+
		1078	+ Logging::GetInstance()->Start(CLIENT_LOG_FILENAME, LogProcessRole::CLIENT, 1);
		1079	+
		1080	+ ASSERT_EQ(FLAGS_log_filename, "config_client");
		1081	+}
		1082	+
		1083	+TEST_F(LoggingTest, TestClientLogNameEnvAfterLoggingRestart)
		1084	+{
		1085	+ Logging::GetInstance()->Start("datasystem_coordinator", LogProcessRole::COORDINATOR, 1);
		1086	+ ResetLoggingForTest();
		1087	+ ScopedEnv logName(LOG_NAME_ENV, "test_client_after_restart");
		1088	+
		1089	+ Logging::GetInstance()->Start(CLIENT_LOG_FILENAME, LogProcessRole::CLIENT, 1);
		1090	+
		1091	+ ASSERT_EQ(FLAGS_log_filename, "test_client_after_restart");
		1092	+}
1060	} // namespace ut	1093	} // namespace ut
1061	} // namespace datasystem	1094	} // namespace datasystem

Mtests/ut/common/util/validator_test.cpp+9-0

	@@ -70,6 +70,15 @@ TEST_F(ValidatorTest, TestIsRegexMatch)
70	EXPECT_FALSE(Validator::IsRegexMatch(simpleIdRe1, "wqeiqwo$"));	70	EXPECT_FALSE(Validator::IsRegexMatch(simpleIdRe1, "wqeiqwo$"));
71	}	71	}
72		72
		73	+TEST_F(ValidatorTest, TestValidateLogName)
		74	+{
		75	+ EXPECT_TRUE(Validator::ValidateLogName("client_log_01"));
		76	+ EXPECT_FALSE(Validator::ValidateLogName(""));
		77	+ EXPECT_FALSE(Validator::ValidateLogName("client-log"));
		78	+ EXPECT_FALSE(Validator::ValidateLogName("client/log"));
		79	+ EXPECT_FALSE(Validator::ValidateLogName("client.log"));
		80	+}
		81	+
73	TEST_F(ValidatorTest, TestValidateHostPortString)	82	TEST_F(ValidatorTest, TestValidateHostPortString)
74	{	83	{
75	EXPECT_TRUE(Validator::ValidateHostPortString("FlagName", ""));	84	EXPECT_TRUE(Validator::ValidateHostPortString("FlagName", ""));

	@@ -424,7 +424,7 @@ Failure-sensitive steps:
424	\| --- \| --- \| --- \| --- \| --- \| --- \| --- \|	424	\| --- \| --- \| --- \| --- \| --- \| --- \| --- \|
425	\| `spdlog` via `ds_spdlog` \| ordinary file and stderr logging, async logger support, rotating sink support \| mature logging backend with async and sink abstractions while preserving a repository-owned macro surface \| keep a fully custom logger, or expose backend implementation details directly \| `MIT` \| pinned to `spdlog-1.12.0` in `bazel/ds_deps.bzl`; upgrade through wrapper and patch review \| repository applies namespace and rotating-sink patches under `third_party/patches/spdlog/*`; retention remains externally managed by `LogManager` \|	425	\| `spdlog` via `ds_spdlog` \| ordinary file and stderr logging, async logger support, rotating sink support \| mature logging backend with async and sink abstractions while preserving a repository-owned macro surface \| keep a fully custom logger, or expose backend implementation details directly \| `MIT` \| pinned to `spdlog-1.12.0` in `bazel/ds_deps.bzl`; upgrade through wrapper and patch review \| repository applies namespace and rotating-sink patches under `third_party/patches/spdlog/*`; retention remains externally managed by `LogManager` \|
426	\| `abseil-cpp` \| failure-signal handling and symbolization for crash-path output \| reduces custom signal-handler code and keeps crash output integrated with stack symbolization \| custom signal handling only \| `Apache-2.0` \| pinned to `abseil-cpp-20250127.1` in `bazel/ds_deps.bzl`; upgrade with crash-path regression checks \| signal-handler behavior is high risk; verify `container.log` output and handler install paths after upgrades \|	426	\| `abseil-cpp` \| failure-signal handling and symbolization for crash-path output \| reduces custom signal-handler code and keeps crash output integrated with stack symbolization \| custom signal handling only \| `Apache-2.0` \| pinned to `abseil-cpp-20250127.1` in `bazel/ds_deps.bzl`; upgrade with crash-path regression checks \| signal-handler behavior is high risk; verify `container.log` output and handler install paths after upgrades \|
427	-\| `RE2` \| log-name validation in `Logging::ValidateLogName(...)` \| safe and predictable regex engine for user-controlled environment input \| manual validation logic \| `BSD-3-Clause` \| pinned to `re2-2024-07-02` in `bazel/ds_deps.bzl`; upgrade with validation tests \| repository carries an absl-related patch for Bazel integration \|	427	+\| `RE2` \| log-name validation in `Validator::ValidateLogName(...)` \| safe and predictable regex engine for user-controlled configuration and environment input \| manual validation logic \| `BSD-3-Clause` \| pinned to `re2-2024-07-02` in `bazel/ds_deps.bzl`; upgrade with validation tests \| repository carries an absl-related patch for Bazel integration \|
428		428
429	TransferEngine reuses the same patched `libds-spdlog.so` ABI but constructs its file and console loggers directly without registering them globally. The private facade owns severity files, stderr thresholds, verbosity controls, flush intervals, rotation size, and default `/tmp`-style directory selection under the `TRANSFER_ENGINE_*` namespace. This boundary is required because `_transfer_engine.so` can coexist with other native Python modules that bring independent logging runtimes.	429	TransferEngine reuses the same patched `libds-spdlog.so` ABI but constructs its file and console loggers directly without registering them globally. The private facade owns severity files, stderr thresholds, verbosity controls, flush intervals, rotation size, and default `/tmp`-style directory selection under the `TRANSFER_ENGINE_*` namespace. This boundary is required because `_transfer_engine.so` can coexist with other native Python modules that bring independent logging runtimes.
430		430
	@@ -442,6 +442,8 @@ TransferEngine reuses the same patched `libds-spdlog.so` ABI but constructs its
442	\| `DATASYSTEM_CLIENT_LOG_DIR` \| environment \| client-only override \| changes client log destination directory \| can redirect output unexpectedly or fail on bad permissions \|	442	\| `DATASYSTEM_CLIENT_LOG_DIR` \| environment \| client-only override \| changes client log destination directory \| can redirect output unexpectedly or fail on bad permissions \|
443	\| `DATASYSTEM_CLIENT_LOG_NAME` \| environment \| client-only override validated by RE2 \| changes client ordinary log base name \| unsafe naming would break file handling if validation changed \|	443	\| `DATASYSTEM_CLIENT_LOG_NAME` \| environment \| client-only override validated by RE2 \| changes client ordinary log base name \| unsafe naming would break file handling if validation changed \|
444	\| `DATASYSTEM_CLIENT_ACCESS_LOG_NAME` \| environment \| client-only override validated by RE2 \| changes client access-log base name \| can break downstream file discovery assumptions \|	444	\| `DATASYSTEM_CLIENT_ACCESS_LOG_NAME` \| environment \| client-only override validated by RE2 \| changes client access-log base name \| can break downstream file discovery assumptions \|
		445	+\| `KVClientConfig::LogName` \| client initialization config \| explicit value takes priority over `DATASYSTEM_CLIENT_LOG_NAME`; explicit empty selects the default name \| changes client ordinary log base name \| config presence must remain distinct from an empty value \|
		446	+\| `KVClientConfig::AccessLogName` \| client initialization config \| explicit value takes priority over `DATASYSTEM_CLIENT_ACCESS_LOG_NAME`; explicit empty selects the default name \| changes client access-log base name \| config presence must remain distinct from an empty value \|
445	\| `DATASYSTEM_LOG_MONITOR_ENABLE` \| environment \| client-only override \| enables or disables client monitor logging \| may create client and server observability mismatch \|	447	\| `DATASYSTEM_LOG_MONITOR_ENABLE` \| environment \| client-only override \| enables or disables client monitor logging \| may create client and server observability mismatch \|
446	\| `log_only_write_info_file` / `DATASYSTEM_LOG_ONLY_WRITE_INFO_FILE` \| gflag or client-only env override \| default `true` \| INFO files always receive all severities; `true` suppresses additional WARNING/ERROR files, while `false` restores dedicated WARNING/ERROR files and severity fanout \| changes file discovery assumptions and disk usage \|	448	\| `log_only_write_info_file` / `DATASYSTEM_LOG_ONLY_WRITE_INFO_FILE` \| gflag or client-only env override \| default `true` \| INFO files always receive all severities; `true` suppresses additional WARNING/ERROR files, while `false` restores dedicated WARNING/ERROR files and severity fanout \| changes file discovery assumptions and disk usage \|
447	\| `TRANSFER_ENGINE_LOG_DIR` \| environment \| unset; then `TEST_TMPDIR`, `TMPDIR`, `TMP`, `/tmp`, `.` \| changes TransferEngine severity-log destination \| invalid or unwritable paths fall through to stderr fallback and reduce persisted diagnostics \|	449	\| `TRANSFER_ENGINE_LOG_DIR` \| environment \| unset; then `TEST_TMPDIR`, `TMPDIR`, `TMP`, `/tmp`, `.` \| changes TransferEngine severity-log destination \| invalid or unwritable paths fall through to stderr fallback and reduce persisted diagnostics \|
	@@ -546,7 +548,7 @@ TransferEngine reuses the same patched `libds-spdlog.so` ABI but constructs its
546	- most logging behavior remains configurable through gflags and client environment overrides;	548	- most logging behavior remains configurable through gflags and client environment overrides;
547	- `log_monitor` is the main kill switch for structured monitor logging.	549	- `log_monitor` is the main kill switch for structured monitor logging.
548	- Attack or fault scenarios the module should continue to tolerate:	550	- Attack or fault scenarios the module should continue to tolerate:
549	- - malformed client-provided log names rejected by `ValidateLogName(...)`;	551	+ - malformed non-empty client-provided log names rejected by `Validator::ValidateLogName(...)`;
550	- async trace discontinuity at one callsite should not corrupt unrelated threads;	552	- async trace discontinuity at one callsite should not corrupt unrelated threads;
551	- crash handler should still emit direct file output when normal logger sinks are unhealthy.	553	- crash handler should still emit direct file output when normal logger sinks are unhealthy.
552		554
	@@ -556,7 +558,8 @@ TransferEngine reuses the same patched `libds-spdlog.so` ABI but constructs its
556	- the module is infrastructure only and must not become an authority for permission decisions.	558	- the module is infrastructure only and must not become an authority for permission decisions.
557	- Input validation requirements:	559	- Input validation requirements:
558	- `Context::SetTraceId(...)` validates trace-prefix characters and length;	560	- `Context::SetTraceId(...)` validates trace-prefix characters and length;
559	- - `Logging::ValidateLogName(...)` only accepts `[a-zA-Z0-9_]*` for client-supplied log names.	561	+ - `Validator::ValidateLogName(...)` only accepts `[a-zA-Z0-9_]+` for client-supplied log names; optional configuration
		562	+ adapters handle an explicit empty value before invoking the common validator.
560	- Data sensitivity or privacy requirements:	563	- Data sensitivity or privacy requirements:
561	- request and response metadata logged by `AccessRecorder` may contain sensitive business context, so any schema expansion requires review for secrets, personal data, or credentials.	564	- request and response metadata logged by `AccessRecorder` may contain sensitive business context, so any schema expansion requires review for secrets, personal data, or credentials.
562	- Secrets, credentials, or key-management requirements:	565	- Secrets, credentials, or key-management requirements:

	@@ -78,18 +78,21 @@ KVClientConfig
78	.. cpp:function:: Builder &LogName(const std::string &name)	78	.. cpp:function:: Builder &LogName(const std::string &name)
79		79
80	设置客户端日志文件名（不含目录和扩展名）。对应内部配置项 ``log_filename``。	80	设置客户端日志文件名（不含目录和扩展名）。对应内部配置项 ``log_filename``。
81	- 不允许为空；允许字符为英文字母、数字及路径字符集中的符号。默认值为 ``ds_client``。	81	+ 非空时仅允许英文字母、数字和下划线；为空时忽略环境变量
		82	+ ``DATASYSTEM_CLIENT_LOG_NAME`` 并使用默认基名 ``ds_client``。默认情况下实际基名为
		83	+ ``ds_client_<pid>``，启用 ``LogWithoutPid`` 后为 ``ds_client``。
82		84
83	.. cpp:function:: Builder &LogWithoutPid(bool enable)	85	.. cpp:function:: Builder &LogWithoutPid(bool enable)
84		86
85	设置客户端日志文件名是否不带 pid。对应环境变量 ``DATASYSTEM_CLIENT_LOG_WITHOUT_PID``。	87	设置客户端日志文件名是否不带 pid。对应环境变量 ``DATASYSTEM_CLIENT_LOG_WITHOUT_PID``。
86	为 ``true`` 时输出为 ``ds_client.log``、``ds_client_access.log``；为 ``false`` 时恢复带	88	为 ``true`` 时输出为 ``ds_client.log``、``ds_client_access.log``；为 ``false`` 时恢复带
87	- ``<pid>`` 后缀的命名，适合多 client 进程同时运行且不希望覆盖日志的场景。默认值为 ``true``。	89	+ ``<pid>`` 后缀的命名，适合多 client 进程同时运行且不希望覆盖日志的场景。默认值为 ``false``。
88		90
89	.. cpp:function:: Builder &AccessLogName(const std::string &name)	91	.. cpp:function:: Builder &AccessLogName(const std::string &name)
90		92
91	设置客户端访问日志文件名（不含目录和扩展名）。非空时仅允许英文字母、数字和下划线。	93	设置客户端访问日志文件名（不含目录和扩展名）。非空时仅允许英文字母、数字和下划线。
92	- 默认值为 ``ds_client_access``。	94	+ 为空时忽略环境变量 ``DATASYSTEM_CLIENT_ACCESS_LOG_NAME`` 并使用默认基名
		95	+ ``ds_client_access``。
93		96
94	.. cpp:function:: Builder &MinLogLevel(int level)	97	.. cpp:function:: Builder &MinLogLevel(int level)
95		98

	@@ -1240,7 +1240,7 @@ dscli query route \
1240	\| watch_event_dispatch_thread \| int \| `4` \| 否 \| Coordinator 分发 Watch 事件的线程数 \|	1240	\| watch_event_dispatch_thread \| int \| `4` \| 否 \| Coordinator 分发 Watch 事件的线程数 \|
1241	\| rpc_thread_num \| int \| `64` \| 否 \| Coordinator RPC 服务线程数 \|	1241	\| rpc_thread_num \| int \| `64` \| 否 \| Coordinator RPC 服务线程数 \|
1242	\| log_dir \| string \| `"./datasystem/logs"` \| 否 \| Coordinator 日志目录 \|	1242	\| log_dir \| string \| `"./datasystem/logs"` \| 否 \| Coordinator 日志目录 \|
1243	-\| log_filename \| string \| `"datasystem_coordinator"` \| 否 \| Coordinator 日志文件名前缀 \|	1243	+\| log_filename \| string \| `"datasystem_coordinator"` \| 否 \| Coordinator 日志文件名前缀，非空时仅允许英文字母、数字和下划线 \|
1244	\| minloglevel \| int \| `0` \| 是 \| 最低日志级别，低于该级别的日志不会被记录 \|	1244	\| minloglevel \| int \| `0` \| 是 \| 最低日志级别，低于该级别的日志不会被记录 \|
1245	\| log_async_queue_size \| int \| `2048` \| 否 \| 异步日志消息队列最大容量 \|	1245	\| log_async_queue_size \| int \| `2048` \| 否 \| 异步日志消息队列最大容量 \|
1246	\| max_log_size \| int \| `400` \| 否 \| 单个日志文件最大大小，单位为 MB \|	1246	\| max_log_size \| int \| `400` \| 否 \| 单个日志文件最大大小，单位为 MB \|
	@@ -1389,7 +1389,7 @@ Coordinator 按该成员列表启动 Raft 选主。启用选主后，`coordinato
1389	\| log_async_queue_size \| int \| `65536` \| 否 \| 异步日志的消息队列最大容量（消息条数） \|	1389	\| log_async_queue_size \| int \| `65536` \| 否 \| 异步日志的消息队列最大容量（消息条数） \|
1390	\| log_compress \| bool \| `false` \| 是 \| 控制是否启用日志压缩功能。启用时，历史日志将自动压缩为gzip格式存储 \|	1390	\| log_compress \| bool \| `false` \| 是 \| 控制是否启用日志压缩功能。启用时，历史日志将自动压缩为gzip格式存储 \|
1391	\| logbufsecs \| int \| `10` \| 否 \| 日志消息最多缓冲时长（以秒为单位） \|	1391	\| logbufsecs \| int \| `10` \| 否 \| 日志消息最多缓冲时长（以秒为单位） \|
1392	-\| log_filename \| string \| `""` \| 否 \| 日志前缀名，当值为空时前缀名为 `datasystem_worker` \|	1392	+\| log_filename \| string \| `""` \| 否 \| 日志前缀名，非空时仅允许英文字母、数字和下划线；为空时前缀名为 `datasystem_worker` \|
1393	\| log_retention_day \| int \| `0` \| 否 \| 日志保留天数，当该值大于0时，最后修改时间早于 `logRetentionDay` 的日志文件将会被删除；当该值为0时表示禁用该功能 \|	1393	\| log_retention_day \| int \| `0` \| 否 \| 日志保留天数，当该值大于0时，最后修改时间早于 `logRetentionDay` 的日志文件将会被删除；当该值为0时表示禁用该功能 \|
1394	\| max_log_file_num \| int \| `5` \| 是 \| 最大日志文件个数，当日志文件个数超过该值时，会将最旧的日志文件删除，通过日志滚动机制保证日志文件最大个数小于等于该值 \|	1394	\| max_log_file_num \| int \| `5` \| 是 \| 最大日志文件个数，当日志文件个数超过该值时，会将最旧的日志文件删除，通过日志滚动机制保证日志文件最大个数小于等于该值 \|
1395	\| max_log_size \| int \| `400` \| 否 \| 单个日志文件最大大小（以MB为单位） \|	1395	\| max_log_size \| int \| `400` \| 否 \| 单个日志文件最大大小（以MB为单位） \|