- Use synchronous pool removal in desktop disconnect_db to prevent
race condition when user quickly disconnects and reconnects
- Rename recordMetadataLoadError to recordConnectionLostError and
export it for cross-store reuse
- Sync query execution errors back to connection state via
recordConnectionLostError in queryStore catch blocks
- Add backend check_connection_health method (dbx-core) with Tauri
command and web route for proactive pool verification
- Add frontend checkConnectionHealth API (tauri.ts/http.ts/api.ts)
- Enhance ensureConnected to verify backend pool health before
assuming connection is valid, auto-reconnect if stale
Root cause:
- AgentDriverClient::kill() is called during Drop to clean up child processes
- Old implementation called child.wait() and ignored errors with let _
- When wait() fails (e.g., process already killed by signal), child is not reaped and becomes zombie
Fix:
- Use try_wait() non-blocking polling, up to 100ms timeout
- Avoid long blocking in Drop during async cleanup
- Still perform final blocking wait() as fallback after timeout
- Log all wait errors for debugging
Impact:
- All JDBC agents (Oracle/DB2/DM8/XuguDB etc.)
- Prevents zombie process accumulation during long-running sessions
Disable COM_RESET_CONNECTION for single-connection session pools
(max_connections == 1) so that MySQL session state such as TEMPORARY
TABLEs is not cleared when a connection is returned to the pool.
Closes#1509
* fix(redis):format INFO output as plain text in cluster mode
* fix(client): remove default JRE download on first launch
* fix(grid): keep row number colors stable
* feat(grid): add pinned query result history
* fix(grid): limit field count badge to data tabs
* fix(settings): persist sidebar page size
* fix(redis): resolve large integer precision loss in JSON tree view
---------
Co-authored-by: Scars <yss2813483011xxl@gmail.com>
Co-authored-by: t8y2 <1156263951@qq.com>
t8y2