Description
In the current implementation, all client rpc use same channel pool.
It can cause a channel closed exception. we need both shared pool and new pool.
In details,
- Fix the TajoAsyncDispatcher hang
- Fix the Fetcher timeout
- Fix the TaskRunner thread leak
- Fix the client rpc reconnecting
- Fix the unittest failure(No available resources)
- Improve RPC thread sharing
Attachments
Attachments
Issue Links
- blocks
-
TAJO-510 Query session timeout too short
- Resolved