proxysql

Commit Graph

Author	SHA1	Message	Date
Rene Cannao	e762f91cda	MCP stats: add in-memory processlist filtering and PgSQL concurrency stress TAP Replace direct stats-schema reads in MCP show_processlist with in-memory processlist access via SQL3_Processlist query options. Add processlist query option types in proxysql_admin interfaces and implement filtering, sorting, and pagination post-processing in both MySQL and PgSQL thread implementations. Introduce configurable MCP cap mcp_stats_show_processlist_max_rows (default 200, max 1000), wire it through MCP variable loading, and expose it in default config. Expand Stats_Tool_Handler processlist handling to parse filter/sort arguments, enforce cap metadata, and return consistent MCP payload fields from in-memory snapshots. Add extensive doxygen documentation in the touched headers and source blocks to describe API contracts, filtering behavior, and runtime constraints. Add new TAP test mcp_pgsql_concurrency_stress-t that generates sustained concurrent PgSQL traffic (simple reads, read/write table workload, randomized pg_sleep) while polling MCP show_processlist and show_queries in parallel with filter and ordering assertions. Validation performed locally: mcp_pgsql_concurrency_stress-t (31 assertions) and mcp_show_queries_topk-t (12 assertions).	4 weeks ago
Rene Cannao	70b61a86e3	MCP stats: implement in-memory show_queries Top-K path Replace stats table reads in show_queries with Query Processor in-memory aggregation to avoid stale/empty admin-db snapshots.\n\nIntroduce Top-K digest filtering primitives and wire them into Stats_Tool_Handler so filtering/sorting/limit are executed against live digest memory. Add runtime MCP variable mcp-stats_show_queries_max_rows with validation, defaulting, and configuration plumbing, plus cap metadata in the tool response.\n\nEnforce configurable hard cap semantics while preserving deterministic ordering and pagination behavior for MCP clients. Add internal DEBUG validator path (PROXYSQLTEST 56) in admin tests to stress generated digest data and validate the Top-K/filter pipeline against in-memory state.\n\nKeep the validator callable from startup/testing workflows without depending on MCP linkage. Add TAP coverage in test/tap/tests/mcp_show_queries_topk-t.cpp for:\n- MCP endpoint setup and reachability\n- digest generation + optional DEBUG internal validator\n- payload validation and cap metadata assertions\n- descending count order validation\n- match_digest_text filter verification\n\nMake payload parsing backward-compatible in TAP (direct tool result object and legacy wrapped success/result payloads).	4 weeks ago
Rene Cannao	c126e63a23	mcp stats: refresh stats tables under admin mutex before direct reads Problem MCP stats tools were querying stats.* tables directly through admindb. Unlike admin-session SQL, that path bypassed GenericRefreshStatistics() and could return stale or empty data for runtime-populated tables. What changed - Updated Stats_Tool_Handler::execute_admin_query() to mirror admin-session semantics for stats reads: - acquire GloAdmin->sql_query_global_mutex - optionally invoke ProxySQL_Admin::GenericRefreshStatistics(sql, ..., false) - execute admindb statement - release sql_query_global_mutex - Added strict input validation and explicit lock/unlock error reporting. - Added refresh_before_query parameter (default true) to avoid duplicate refresh passes for secondary count queries. - Switched secondary COUNT(*) calls in show_processlist/show_queries/ show_errors/show_query_rules to refresh_before_query=false. Documentation - Expanded Doxygen in include/Stats_Tool_Handler.h for execute_admin_query() to document locking, refresh behavior, and performance tradeoff. - Added detailed Doxygen above the execute_admin_query() implementation in lib/Stats_Tool_Handler.cpp. Testing - Added TAP integration test test/tap/tests/mcp_stats_refresh-t.cpp. - The test injects a synthetic stale marker row into stats.stats_mysql_global, calls /mcp/stats show_status, and verifies the marker disappears after refresh-before-read. - Added helper-level Doxygen in the test for setup/parsing behavior. Build verification - Compiled lib object: make -C lib obj/Stats_Tool_Handler.oo - Compiled TAP target: make -C test/tap/tests mcp_stats_refresh-t - Direct runtime execution of the TAP test in this workspace could not fully run because no local admin listener was available (connection to 127.0.0.1:6032 failed).	4 weeks ago
Rene Cannao	33fce1ca35	MCP stats: address review findings and add detailed docs	4 weeks ago
René Cannaò	762afaa44f	Merge branch 'v3.0' into v4.0-mcp-stats Signed-off-by: René Cannaò <rene@proxysql.com>	1 month ago
René Cannaò	10ff1322ba	Merge pull request #5386 from sysown/v3.0-MCP_multi MCP query+harvest hardening: HGM routing, weighted targets, pgsql harvest fix, TAP updates	1 month ago
Rene Cannao	253d098653	security: fix SQL injection vulnerabilities in SQLite catalog queries Address Gemini Code Assist review feedback on PR #5386: 1. Discovery_Schema::resolve_run_id - Replace string concatenation with replace_str escaping with proper parameterized queries using sqlite3_prepare_v2 and sqlite3_bind_* functions. 2. PgSQL_Static_Harvester - Remove unsafe escape_sql_string() function and replace its usages with: - New lookup_object_id() helper using parameterized queries - Parameterized UPDATE statements in harvest_view_definitions() These changes ensure complete protection against SQL injection for the catalog database, following the pattern already used in create_run().	1 month ago
Rene Cannao	2fa1401b05	MCP query+harvest hardening: HGM-based target routing, weighted backend selection, pgsql schema fix, and TAP robustness This squashed commit consolidates the recent MCP stabilization work across query execution and static-discovery tests. 1) Query endpoint routing/executability now uses Hostgroup Manager runtime snapshots - File: lib/Query_Tool_Handler.cpp - Removed dependency on direct SQL probes against runtime_mysql_servers/runtime_pgsql_servers from inside the MCP query handler path. - Target backend resolution now reuses the same Hostgroup Manager runtime snapshot APIs that feed Admin runtime server visibility: - MyHGM->dump_table_mysql("mysql_servers") - PgHGM->dump_table_pgsql("pgsql_servers") - Non-executable diagnostics now summarize statuses from HGM snapshots instead of ad hoc runtime-table SQL. 2) Backend endpoint selection changed to weighted random - File: lib/Query_Tool_Handler.cpp - Selection remains constrained by: - matching target hostgroup_id - ONLINE backend status - Candidate choice now uses random selection with linear probability by weight, replacing deterministic ordering. 3) Read-only validation and guardrail adjustments - File: lib/Query_Tool_Handler.cpp - Fixed false negatives for valid read-only queries such as SELECT literals (no FROM clause required). - Added additional dangerous-query patterns for file-write variants: - INTO OUTFILE - INTO DUMPFILE 4) PostgreSQL static harvester schema query fix - File: lib/PgSQL_Static_Harvester.cpp - Fixed harvest_schemas() failure on PostgreSQL caused by querying non-existent information_schema.schemata.default_collation_name. - Replaced with PostgreSQL-safe projection while preserving the insert_schema() column contract. 5) TAP reliability improvements for MCP test phases - File: test/tap/tests/mcp_rules_testing/test_phase11_pgsql_target.sh - Isolated phase state by clearing MCP rules before setup. - Added verbose response prints for T11.1-T11.4. - Added LOGENTRY phase start/end markers and EXIT trap for guaranteed end watermark. - Replaced fragile grep/escaping checks with robust string assertions. - Fixed NULL-concat stats assertions with COALESCE(username,'') and COALESCE(target_id,''). - File: test/tap/tests/test_mcp_static_harvest-t.sh - Hardened escaped-JSON parsing in MCP response checks (run_id extraction + key/value assertions). - Reworked target/protocol/object assertions to avoid transport-format false negatives. - Corrected cross-target negative assertion to check MCP isError semantics. Observed result in user verification - Phase 11 pgsql target/rules/stats checks pass. - Static harvest phase-A checks pass for mixed mysql/pgsql target_id flow. This commit intentionally keeps runtime ONLINE-only target executability semantics and improves observability/diagnostic quality for future MCP routing/test failures.	1 month ago
Rene Cannao	9685cdaa4b	mcp discovery: enforce target-scoped run model and add protocol-aware static harvesting (mysql+pgsql) Refactor the MCP query/discovery stack to remove prototype-era MySQL-only assumptions and make discovery/catalog semantics explicitly target-scoped. This commit is intentionally not backward compatible with legacy single-target catalog metadata. Key implementation details: - Discovery_Schema: - runs table now stores target_id, protocol, and server_version - resolve_run_id() now requires target_id and resolves schemas within target scope - create_run() now records target/protocol/server version - added legacy schema detection + destructive catalog table rebuild when old runs layout is found - fixed resultset lifetime/escaping issues in target-aware run resolution - New PostgreSQL static harvester: - added PgSQL_Static_Harvester class and build wiring - harvests schemas/objects/columns/indexes/fks/view definitions into Discovery_Schema - maps pg metadata to object_id-based insert APIs used by Discovery_Schema - Query_Tool_Handler: - constructor simplified to catalog_path only (no mcp-mysql_* runtime ctor deps) - discovery.run_static now requires target_id and dispatches protocol-aware harvester at runtime - catalog., agent.run_start, and llm. tool contracts now require target_id when resolving run_id - all run_id resolution call sites switched to resolve_run_id(target_id, ...) - list_schemas catalog query now scoped by target_id via runs join - digest logging path now resolves run_id in target context - ProxySQL_MCP_Server: - updated Query_Tool_Handler construction to new signature - Docs updated: - added explicit target_id scoping guidance for discovery/catalog/agent/llm workflows - clarified protocol-aware routing and noted legacy examples still present in large script docs Build validation: - Recompiled changed objects with PROXYSQLGENAI=1: - Query_Tool_Handler.oo - Discovery_Schema.oo - PgSQL_Static_Harvester.oo - ProxySQL_MCP_Server.oo	1 month ago
Rene Cannao	7e54883ba5	mcp query debugging: surface rule-id context for blocks and add backend failure SQL details Observed confusion: - MCP failures like 'Error 100' / 'Error 101' were visible at endpoint level but gave no direct clue whether they came from query-rule blocks or backend execution failures. - Connection/firewall-style failures also lacked enough target/query context in lower-level logs. Changes in this commit: 1) Log MCP query-rule block/OK matches with rule id and context - In Discovery_Schema::evaluate_mcp_query_rules(): - when error_msg rule action is applied, log: rule_id, tool, target_id, schema, error_msg, query preview. - when OK_msg rule action is applied, log: rule_id, tool, target_id, schema, ok_msg. - This makes it explicit which runtime_mcp_query_rules row produced the returned error. 2) Add detailed backend execution error logs in Query_Tool_Handler - For MySQL and PostgreSQL execution paths (with and without schema switching), failures now log: - target_id - schema (when applicable) - backend error text - SQL query text - Added context for PG search_path failures too. - This disambiguates rule-block errors from actual backend connectivity/authorization/firewall failures. Behavioral notes: - API responses are unchanged (to avoid breaking tests expecting exact messages). - Logging is now sufficiently descriptive to trace each failure to either: - a specific MCP rule id/action, or - a concrete backend execution error. Validation: - Recompiled successfully: - lib/obj/Discovery_Schema.oo - lib/obj/Query_Tool_Handler.oo	1 month ago
Rene Cannao	8fbd570c79	mcp variables: stop writing runtime_global_variables during LOAD MCP VARIABLES TO RUNTIME Issue: - MCP variable load path was writing rows into runtime_global_variables. - For MCP this is incorrect and misleading; runtime_global_variables should not be used as a write target in this flow. - Logs showed explicit INSERT statements into runtime_global_variables for mcp-* entries. Fix implemented: 1) Removed runtime_global_variables population from MCP database->runtime flow - In flush_mcp_variables___database_to_runtime(), removed the block that called: flush_mcp_variables___runtime_to_database(..., runtime=true, ...) - LOAD MCP VARIABLES TO RUNTIME now applies mcp-* variables to GloMCPH and reloads MCP server state, without runtime_global_variables writes. 2) Removed runtime_global_variables writes from MCP runtime->database helper - In flush_mcp_variables___runtime_to_database(): - removed DELETE FROM runtime_global_variables WHERE variable_name LIKE 'mcp-%' - removed INSERT INTO runtime_global_variables(...) per variable - Function now persists MCP variables only to global_variables (main/disk paths), which is the intended save behavior. 3) Removed temporary MCP info-level debug noise related to runtime_global_variables inserts - Deleted MCP info logs that printed per-variable INSERT SQL into runtime_global_variables. Build validation: - Recompiled successfully: - lib/obj/Admin_FlushVariables.oo Expected result after this commit: - Executing LOAD MCP VARIABLES TO RUNTIME should no longer emit any INSERT/DELETE against runtime_global_variables for mcp-* variables. - MCP runtime behavior remains: variables are applied in-memory and MCP server reload logic is triggered via load_mcp_server().	1 month ago
Rene Cannao	f15460348c	mcp query diagnostics: include runtime hostgroup status breakdown when target lacks ONLINE backend Adds targeted debug context for non-executable MCP targets. When Query_Tool_Handler reports that a target has no ONLINE backend, it now appends a status summary read from the corresponding runtime server table for that hostgroup: - MySQL: runtime_mysql_servers - PostgreSQL: runtime_pgsql_servers Example diagnostics now include status aggregates such as: - ONLINE=0, SHUNNED=1 - OFFLINE_SOFT=1 - no rows in runtime_*_servers for hostgroup X This closes the observability gap where users saw 'target non-executable' but could not immediately tell why the hostgroup failed ONLINE eligibility. Validation: - Recompiled lib/obj/Query_Tool_Handler.oo successfully.	1 month ago
Rene Cannao	f4bc1943fb	mcp query diagnostics: restore strict ONLINE requirement and explain target non-executable failures This change does two things: 1) Reverts executable-target policy to strict ONLINE backend status - Query target resolution in Query_Tool_Handler is restored to require UPPER(status)='ONLINE' in runtime_mysql_servers/runtime_pgsql_servers. - This matches expected semantics: non-ONLINE backends should not be considered executable MCP targets. 2) Replaces generic non-executable errors with actionable diagnostics - Added Query_Tool_Handler::format_target_unavailable_error(target_id). - All query-tool paths that previously returned the generic: 'Target is not executable by this handler' now return a reasoned error with context, including: - target_id - protocol - hostgroup_id - auth_profile_id - concrete reason (for example: empty db_username in auth profile, or no ONLINE backend in runtime_*_servers) - Internal logs in get_connection()/get_pgsql_connection() now emit the same detailed reason. 3) Improves MCP endpoint failure logging with SQL/target context - In MCP_JSONRPC_Resource::handle_tools_call(), failed tool calls now log: - endpoint/tool/error - full arguments payload (existing) - additional parsed details when present: target_id, schema, sql (trimmed for safety) - This makes it explicit what SQL was requested when a tool call fails before execution. 4) Protocol normalization retained for robustness - MCP_Threads_Handler::load_target_auth_map() lowercases protocol values when loading runtime profile joins. - This avoids protocol-casing drift causing inconsistent routing behavior. Build validation: - Successfully recompiled modified objects: - lib/obj/Query_Tool_Handler.oo - lib/obj/MCP_Endpoint.oo - lib/obj/MCP_Thread.oo Resulting behavior expected on next test run: - If a target is non-executable, the response and logs will state exactly why (ONLINE/backend/auth reason), instead of a generic error. - Failure logs will also include the attempted SQL text (for tool calls that carry sql arguments), clarifying whether a backend query was actually executed or blocked before execution.	1 month ago
Rene Cannao	49f811a638	mcp query: stop misclassifying reachable targets as non-executable Observed failure: - MCP tools received valid target_id values, but /mcp/query returned 'Target is not executable by this handler'. - This prevented rule evaluation, digest accounting, and hit counters from advancing, causing multiple TAP phases to fail. Root cause addressed in this commit: - Target endpoint resolution required runtime server status to be exactly ONLINE. - In practice, reachable backends can temporarily appear in other statuses (for example monitor-related transitions), while MCP direct connections with profile credentials are still valid. - Protocol casing inconsistencies could also lead to fragile routing behavior when profile data is consumed. Changes: 1) Relax backend eligibility during endpoint resolution - In Query_Tool_Handler::refresh_target_registry(), endpoint discovery now accepts any server status except OFFLINE_HARD. - ONLINE is still preferred deterministically via ORDER BY, but non-ONLINE candidates remain eligible for MCP execution. 2) Normalize protocol values to lowercase - In MCP_Threads_Handler::load_target_auth_map(), target protocol is normalized to lowercase when loaded from runtime profiles. - In Query_Tool_Handler::refresh_target_registry(), protocol is normalized again defensively before routing decisions. 3) Improve diagnostics for target executability - Added explicit warnings when: - a target resolves to a backend but has empty db_username, - a target has no eligible backend for its hostgroup/protocol. - These logs make it clear why a target is being marked non-executable. Expected behavioral impact: - MCP query tools should execute for valid target_id entries even when monitor status is not strictly ONLINE. - Query-rules blocking/rewriting/OK actions should be exercised again. - stats_mcp_query_rules hits and stats_mcp_query_digest counters should resume incrementing once queries run. Validation performed: - Recompiled modified objects successfully: - lib/obj/Query_Tool_Handler.oo - lib/obj/MCP_Thread.oo	1 month ago
Rene Cannao	6a788e48c4	mcp: make /mcp/query self-healing when targets/backends appear after startup Problem addressed: - MCP query endpoint could stay unusable with 'Tool Handler not initialized' after restart/reload flows. - This was triggered when Query_Tool_Handler could not build an executable pool at init time (for example profiles loaded before ONLINE servers), leaving runtime commands like 'LOAD MCP QUERY RULES FROM MEMORY' blocked behind a NULL query tool handler. - Users had to manually toggle MCP enablement to recover, which is the opposite of expected self-healing behavior. What this commit changes: 1) Query_Tool_Handler pool init is now resilient and idempotent - init_connection_pool() now starts with close() so reinitialization fully resets stale mysql/pgsql pool state before rebuilding from runtime profile + server tables. - If no executable targets are available, init_connection_pool() now returns success with a warning instead of hard failure. This allows the query tool handler (and /mcp/query endpoint) to stay initialized even before backends are ready. 2) Lazy auto-rebuild on first query usage - get_connection() and get_pgsql_connection() now: - refresh target registry before resolution, - attempt to use an existing pooled connection, - if unavailable, trigger a full pool rebuild (init_connection_pool()) and retry once. - This provides automatic recovery when hostgroups/servers/profiles are loaded or changed after MCP startup, without requiring manual MCP disable/enable. 3) Admin runtime rule load path attempts MCP recovery - load_mcp_query_rules_to_runtime() now detects NULL query tool handler and calls load_mcp_server() once before failing. - This turns a hard, immediate admin error into a self-recovery attempt consistent with MCP runtime semantics. Behavioral impact: - /mcp/query endpoint remains online even when there are temporarily zero executable targets. - As soon as compatible runtime targets/backends exist, run_sql_readonly/explain_sql can recover automatically on demand. - LOAD MCP QUERY RULES TO/FROM RUNTIME no longer fails immediately on first NULL handler condition; it retries after MCP server recovery. Validation performed: - Recompiled modified objects successfully: - lib/obj/Query_Tool_Handler.oo - lib/obj/ProxySQL_Admin.oo - Full TAP runtime test execution is not possible in this sandbox due blocked local TCP socket creation; validation should be run in the normal test environment where ProxySQL/MySQL/PGSQL are reachable.	1 month ago
Rene Cannao	998bd82387	MCP TAP startup: fix tool-handler initialization order, improve MCP PROFILES observability, and seed monitor users This change fixes recurring MCP TAP failures where `/mcp/query` returned: Tool Handler not initialized for endpoint: query and where backend monitor auth failures flooded logs. Problem summary - MCP server startup can occur before runtime target/auth profiles and backend server mappings are loaded. - If that happens, Query_Tool_Handler initialization sees no executable targets and remains NULL. - MCP endpoint resources bind the handler pointer at creation time, so a NULL query handler at startup breaks `/mcp/query` until server restart. Code changes 1) Add explicit admin command logging for MCP PROFILES commands - Added `Received <command>` logging in the MCP PROFILES command block, matching behavior of other admin command handlers. - File: `lib/Admin_Handler.cpp` 2) Trigger MCP server refresh after `LOAD MCP PROFILES TO RUNTIME` - After copying profiles into runtime and rebuilding target/auth map, call `ProxySQL_Admin::load_mcp_server()`. - This allows MCP to self-heal when profiles become available after initial startup. - File: `lib/Admin_Handler.cpp` 3) Restart MCP server when query handler is missing - Extended `ProxySQL_Admin::load_mcp_server()` restart checks to include: - running server + `query_tool_handler == NULL` - Restart reason now includes tool handler initialization mismatch. - File: `lib/ProxySQL_Admin.cpp` 4) Fix TAP configurator load order to avoid early MCP startup - Reordered `test/tap/tests/mcp_rules_testing/configure_mcp.sh` runtime sequence: - `LOAD MYSQL SERVERS TO RUNTIME` - `LOAD PGSQL SERVERS TO RUNTIME` (best effort) - `LOAD MCP PROFILES TO RUNTIME` - `LOAD MCP VARIABLES TO RUNTIME` (last) - This ensures MCP starts only after routing/auth context is present. 5) Seed monitor credentials in AI local infra pre-hook - Added backend user/role creation for default monitor credentials `monitor/monitor`: - MySQL: create user + monitor-relevant grants - PostgreSQL: create role + `pg_monitor` + DB connect grants - Reduces monitor auth noise in local AI TAP dockerized setup. - File: `test/tap/groups/ai/pre-proxysql.bash` 6) Mark new TAP phase scripts executable - `test_phase10_eval_explain.sh` - `test_phase11_pgsql_target.sh` Expected outcome - MCP query endpoint no longer stays stuck with an uninitialized tool handler after TAP configuration. - MCP query-rules admin commands stop failing due to missing Query_Tool_Handler. - MCP profile command flow is visible in logs for easier debugging. - Local AI TAP infra no longer emits continuous monitor authentication failures for default monitor credentials.	1 month ago
Rene Cannao	af0411bd46	MCP: add target-aware rules/stats tests, explain_sql rule coverage, and AI local docker TAP infra This commit completes end-to-end MCP query-rules validation for multi-target routing and introduces a self-contained TAP infra for the `ai` group that can run both in Jenkins and manually. Main MCP/runtime changes: - Extended MCP query-rule matching context to include both `target_id` and resolved backend `username`. - Added `target_id` column to `mcp_query_rules` and `runtime_mcp_query_rules` table definitions. - Extended `stats_mcp_query_rules` to include `username` and `target_id` alongside `rule_id` and `hits`. - Updated load/save/runtime refresh paths to persist and rehydrate the expanded MCP rule schema. - Wired the rule engine into `explain_sql` so MCP rules apply consistently across `run_sql_readonly` and `explain_sql`. - Included startup-order fix in `src/main.cpp` to initialize MCP/GenAI thread handlers early, preventing startup crashes in PROXYSQLGENAI builds. Test coverage changes: - Updated existing MCP TAP phases to assert target-aware and username-aware behavior: - `test_phase4_stats.sh` - `test_phase6_eval_block.sh` - Added new MCP TAP phases: - `test_phase10_eval_explain.sh` (rule engine coverage for `explain_sql`) - `test_phase11_pgsql_target.sh` (pgsql target routing/rule/stats coverage; graceful skip if no pgsql target is configured) - Updated `test_mcp_query_rules-t.sh` runner to execute new phases. AI group isolated infra (manual + CI compatible): - Added `test/tap/groups/ai/docker-compose.yml` with MySQL 9.0 and PostgreSQL 16 backends. - Added lifecycle scripts: - `docker-compose-init.bash` - `docker-compose-destroy.bash` - `pre-proxysql.bash` - `post-proxysql.bash` - Extended `test/tap/groups/ai/env.sh` with local default ports/credentials/target IDs used by MCP TAP tests. - Added `test/tap/groups/ai/README.md` documenting manual execution flow outside Jenkins. Outcome: - MCP tests now validate routing-aware rule enforcement and stats attribution for both MySQL and PostgreSQL targets. - The `ai` TAP group can be run with an isolated local backend stack without relying on external Jenkins infra repositories.	1 month ago
René Cannaò	28b1114cd9	Merge pull request #5353 from sysown/v3.0_pgsql-prepared-statement-refcount-race-5352 Fix PostgreSQL prepared statement purge crash: use memory_order_acquire instead of memory_order_relaxed	1 month ago
Rene Cannao	013864b36f	MCP: introduce profile-based target/auth routing and unified LOAD/SAVE MCP PROFILES commands This commit introduces the core MCP routing architecture that decouples client-facing target selection from backend connection details. Architecture introduced: - New MCP auth profile table to store backend credentials/policy (`mcp_auth_profiles`). - New MCP target profile table to map opaque `target_id` -> protocol/hostgroup/auth profile (`mcp_target_profiles`). - Runtime mirrors for both profile tables. Behavioral changes: - MCP query tools resolve execution context by `target_id` and route internally by protocol + hostgroup. - Client no longer needs backend-specific connection details; only logical target identifiers. - Unified admin commands for profile lifecycle management: - `LOAD MCP PROFILES TO/FROM MEMORY/RUNTIME/DISK` - `SAVE MCP PROFILES TO/FROM MEMORY/RUNTIME/DISK` Design goals: - Support multi-backend MCP execution (MySQL + PostgreSQL) behind one endpoint. - Provide server-side credential management and cleaner operational model. - Replace legacy one-backend POC configuration with scalable target/auth abstractions. Result: - MCP routing is now dynamic, protocol-agnostic at the client layer, and fully managed through profile tables and unified load/save commands.	1 month ago
René Cannaò	5de836a4c1	Merge pull request #5373 from sysown/v3.0_improve_mysql_monitoring_5256 Improve MySQL Monitor connection management	1 month ago
René Cannaò	3b578db4bf	Merge pull request #5374 from sysown/v3.0-test0213 Fix Valgrind-reported uninitialized memory issues	1 month ago
René Cannaò	83209da487	Merge pull request #5356 from sysown/v3.0-5355 Fix #5355: Add null pointer check in RequestEnd() to prevent use-after-free crash	1 month ago
Rene Cannao	37e72ea3f9	Fix PROXY protocol detection in MySQL_Data_Stream::read_from_net When pkts_recv==0, the code reads 4 bytes expecting a MySQL packet header, but PROXY protocol starts with 'PROXY ' which gets misinterpreted as a MySQL header with a huge packet length (0x504f5250 = ~1.3GB). Add check for 'PROX' prefix before parsing as MySQL header. This prevents passing an incorrect length parameter to recv() that exceeds the buffer size. Fixes #5376	1 month ago
Rene Cannao	11a43bf768	Fix uninitialized max_allowed_pkt in MySQL_Connection constructor Initialize options.max_allowed_pkt to 0 to prevent Valgrind errors when logging handshake debug information.	1 month ago
Rene Cannao	ca4f858b29	Fix uninitialized memory read in pgsql tokenizer Change malloc() to calloc() in get_result_buffer() to zero-initialize the query digest buffer before use. This prevents stage_2_parsing() from reading uninitialized memory when processing query digests. Fixes #5375	1 month ago
Rene Cannao	85952373d9	Fix uninitialized mondb pointer in MySQL_Monitor_State_Data constructor Add mondb = NULL initialization to prevent Valgrind errors when monitor threads access the pointer before it's properly set.	1 month ago
Rahim Kanji	8ca6323eef	Merge remote-tracking branch 'Master/v3.0' into v3.0_pgsql-prepared-statement-refcount-race-5352	1 month ago
Rene Cannao	200fdecb8d	Apply AI agent review fixes to PR #5374 Critical Fixes: - Fix NULL pointer dereferences in MySQL_Logger.cpp: * Add null checks for myds->myconn->parent in log_audit_entry() * Fix cleanup code in log_request() and log_audit_entry() * Change sprintf to snprintf for buffer safety - Remove redundant pthread_rwlock_init in debug.cpp: * filters_rwlock already initialized with PTHREAD_RWLOCK_INITIALIZER * Re-initialization caused undefined behavior per POSIX Major Improvements: - Add bounded wait helper wait_for_glo_mth() in proxysql_utils.h: * Replaces indefinite while loops with 10s timeout * Applied to 26 locations across 4 files Build System: - Require C++17 in Makefile (remove C++11 fallback): * Code uses C++17 structured bindings extensively * Cleanest solution vs adding fallback guards to 27+ locations Compilation fixes: - Add unistd.h include for usleep - Remove conflicting GloMyMon extern declaration Files changed: - Makefile - include/proxysql_utils.h - lib/MySQL_Logger.cpp - lib/debug.cpp - lib/MySQL_Monitor.cpp - lib/ProxySQL_Admin.cpp - lib/ClickHouse_Server.cpp - src/SQLite3_Server.cpp	1 month ago
Rahim Kanji	1cccfb3220	Fix connection cleanup in failure paths	1 month ago
Rene Cannao	941e6b7320	Fix filters_rwlock initialization race condition Add static initialization PTHREAD_RWLOCK_INITIALIZER to filters_rwlock in debug.cpp to prevent race condition where proxy_debug() is called before init_debug_struct() initializes the rwlock. This fixes Valgrind errors in filter_debug_entry() where threads were accessing an uninitialized rwlock.	1 month ago
Rene Cannao	3c85c593cf	Add GloMTH initialization wait to all monitor and server threads Fix race condition where threads were accessing GloMTH before it was fully initialized, causing Valgrind errors in get_variable_int() and related functions. Added wait loop pattern (while GloMTH==NULL { usleep(50000); }) followed by usleep(100000) to ensure GloMTH is fully initialized before use: In lib/MySQL_Monitor.cpp: - monitor_connect(), monitor_ping(), monitor_read_only() - monitor_group_replication(), monitor_galera(), monitor_replication_lag() - monitor_AWS_Aurora_thread_HG(), monitor_aws_aurora() - init_mysql_thread_struct() In src/SQLite3_Server.cpp: - SQLite3_Server_session_handler In lib/ProxySQL_Admin.cpp: - admin_session_handler In lib/ClickHouse_Server.cpp: - ClickHouse_Server_session_handler The pattern ensures threads wait for GloMTH initialization during startup, then check for NULL during shutdown to exit cleanly.	1 month ago
Rene Cannao	aa53d990ed	Add GloMTH initialization wait to all monitor threads Fix race condition where monitor threads were accessing GloMTH before it was fully initialized, causing Valgrind errors in get_variable_int(). Added wait loop pattern (while GloMTH==NULL { usleep(50000); }) followed by usleep(100000) to: - monitor_ping_thread - monitor_read_only_thread - monitor_replication_lag_thread - monitor_group_replication_thread (also moved check before MySQL_Thread creation) - monitor_galera_thread (also moved check before MySQL_Thread creation) This matches the existing pattern in monitor_connect_thread.	1 month ago
Rene Cannao	f4cd34be5c	Fix race condition in monitor_connect_thread Add wait loop for GloMTH initialization before accessing variables. Other monitor threads (MonitorPing, MonitorReadOnly, etc.) already had this pattern, but monitor_connect_thread was missing it. This fixes Valgrind errors about uninitialized values in get_variable_uint16() when strcasecmp reads from memory that hasn't been fully initialized yet.	1 month ago
Rahim Kanji	33ce4217f0	Merge remote-tracking branch 'v3.0' into v3.0_improve_mysql_monitoring_5256	1 month ago
Rahim Kanji	9eac5f6420	Improve async ping mmsd ownership tracking and pool validation	1 month ago
Rene Cannao	09d0dcb808	Fix uninitialized tmp_charset in Data_Stream constructors Initialize tmp_charset=0 in MySQL_Data_Stream and PgSQL_Data_Stream constructors. This field was used uninitialized in process_pkt_handshake_response() when calling proxy_debug(), causing Valgrind errors about uninitialized values in vsnprintf.	1 month ago
Rene Cannao	7c67eb25f1	Fix uninitialized fd and status in Data_Stream constructors Initialize fd=-1 and status=0 in MySQL_Data_Stream and PgSQL_Data_Stream constructors. These fields were not initialized, causing Valgrind errors when GPFC_Statuses2() checked mybe->server_myds->status and fd.	1 month ago
Rene Cannao	4e39297ef9	Fix more Valgrind uninitialized value errors 1. Base_Thread.cpp: Initialize member variables in constructor - curtime, last_move_to_idle_thread_time, epoll_thread, shutdown, mysql_sessions were not initialized, causing uninitialized value use in ProcessAllMyDS_BeforePoll and ProcessAllMyDS_AfterPoll. 2. MySQL_Logger.cpp: Initialize query_ptr, query_len, and buf in MySQL_Event constructor. These fields were used without initialization, causing Valgrind errors in JSON serialization and other operations.	1 month ago
Rene Cannao	fdbea2a32e	Fix uninitialized memory in Command_Counter::_counters Replace _counters{} initialization syntax with explicit memset() to ensure the _counters array is properly zero-initialized. The _counters{} syntax may not reliably zero-initialize arrays in all compiler implementations, causing Valgrind to report use of uninitialized values when add_and_reset() reads from the array. Also remove the safety checks added previously since the root cause is now properly fixed.	1 month ago
Rene Cannao	c2f82b3d4e	Fix more Valgrind uninitialized value errors 1. MySQL_Logger.cpp: Initialize username_len, schemaname_len, and server_len to 0 in MySQL_Event constructor. These fields were uninitialized and used in copy constructor and JSON serialization. 2. debug.cpp: Add default member initializers to DebugLogEntry struct to ensure POD fields are zero-initialized. Move errno save/restore to after early return checks to avoid potential TLS access issues before thread is fully initialized.	1 month ago
Rene Cannao	ee7ee0b387	Fix multiple Valgrind-reported uninitialized memory issues 1. MySQL_Logger.cpp: Add NULL checks for myconn->parent before accessing address/port. Client connections (frontend) have parent=NULL, causing uninitialized value reads when logging. 2. debug.cpp: Add NULL check for GloVars.global.gdbg_lvl before accessing debug level configuration. Prevents use of uninitialized data when proxy_debug_func is called before init_debug_struct() completes. 3. ProxySQL_Admin.cpp: Add NULL checks for debug filter database fields before constructing filter strings. Prevents undefined behavior when constructing std::string from NULL pointers.	1 month ago
Rene Cannao	af58865b06	Fix uninitialized memory in QueryParserArgs.buf Use memset to zero-initialize the entire QueryParserArgs struct instead of individually setting only digest_text and first_comment to NULL. This fixes Valgrind errors about use of uninitialised values in strcasestr() when reading from the uninitialized buf array. Fixes errors like: Use of uninitialised value of size 8 at tolower (ctype.c:46) by strcasestr by MySQL_Connection::ProcessQueryAndSetStatusFlags(char*)	1 month ago
Rene Cannao	cc53ddca23	Merge branch 'v3.0-openssl-fix' into v3.0-test0213	1 month ago
Rene Cannao	52718df11f	Fix memory leak in stats processlist functions - Add missing delete resultset in stats___mysql_processlist - Add missing delete resultset in stats___pgsql_processlist	1 month ago
Rene Cannao	746eca9f7c	More sprintf to snprintf conversions and indentation fixes - Replace remaining sprintf with snprintf in Admin_FlushVariables.cpp - Fix indentation in flush_GENERIC_variables__process__database_to_runtime - Apply same pattern to ProxySQL_Admin.cpp and ProxySQL_Admin_Stats.cpp	1 month ago
Rene Cannao	3e37337877	Use RAII for sqlite3 statements across codebase - Convert prepare_v2() calls to use unique_ptr RAII pattern - Remove manual sqlite3_finalize calls (now handled by RAII) - Update flush_*_variables functions to use new prepare_v2 return pattern - Apply changes to Admin, Cluster, Monitor, Catalog, and other modules	1 month ago
Rene Cannao	335be9ec72	Fix unsafe sprintf and error message handling - Replace sprintf with snprintf in flush_*_variables functions - Add null value safety with safe_val in variable flushing - Fix error message to use send_error_msg_to_client instead of send_ok_msg_to_client	1 month ago
Rene Cannao	61ce0a96e8	Remove redundant sqlite3_finalize calls in pgSQL stats Statements are managed by RAII and auto-finalized	1 month ago
Rene Cannao	afd6dffb05	Fix admin shutdown races and implement graceful teardown - Add coordinated shutdown for PROXYSQL KILL/SHUTDOWN in debug builds - Track active admin client threads with admin_client_threads_active counter - Add shutdown checks to prevent new connections during shutdown - Wait for all admin client threads to finish before cleanup in admin_shutdown() - Handle accept failures gracefully	1 month ago
Rene Cannao	368a9e3a63	Fix log_buffer debug leak during shutdown	1 month ago

1 2 3 4 5 ...

4893 Commits (e762f91cda174ea3e6ab85bbcbdbc457280edf7f)