Details
-
Task
-
Status: Closed (View Workflow)
-
Critical
-
Resolution: Fixed
-
10.3.6-1, 10.4.0-1
Description
MySQL 8.0.0 split the InnoDB buf_pool_t::mutex. MariaDB should do something similar.
Instead of introducing more mutexes or radically changing the latching rules of various buf_pool_t and buf_block_t data members, I think that it is possible to reduce the contention on buf_pool.mutex by other means:
- Move more code to inline functions of buf_pool_t or buf_page_t.
- Reduce the amount of mutex release/reacquire dance in buf0flu.cc and buf0rea.cc.
- Avoid repeated calls to page_id_t::fold() or page_id_t::page_id_t(); use page_id_t directly as loop iterator.
- Move buf_page_t::flush_type to IORequest.
- Split buf_page_io_complete() into separate ‘read completion’ and ‘write completion’ callbacks.
- Avoid holding buf_pool.mutex during buf_pool.page_hash operations. Consider removing the debug field buf_page_t::in_page_hash.
- Split operations on buf_pool.watch[] into two parts. The allocation of buf_pool.watch[] should be only protected by buf_pool.mutex, and the buf_pool.page_hash only by the hash bucket latch.
Attachments
Issue Links
- blocks
-
-
- Closed
-
- causes
-
MDEV-22816 Assertion `node->space == fil_system.sys_space' failed in fil_aio_callback
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
- is duplicated by
-
-
- Closed
-
- relates to
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Open
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
-
-
- Closed
-
Activity
| Field | Original Value | New Value |
|---|---|---|
| Link |
This issue relates to |
| Link |
This issue relates to |
| Assignee | Marko Mäkelä [ marko ] | Thirunarayanan B [ thiru ] |
| Status | Open [ 1 ] | In Progress [ 3 ] |
| Attachment | bug75534-2.patch [ 45258 ] |
| Link |
This issue relates to |
| Sprint | 10.3.6 [ 237 ] |
| Sprint | 10.3.6-1 [ 237 ] | 10.3.6-1, 10.4.0-1 [ 237, 254 ] |
| Rank | Ranked lower |
| NRE Projects | RM_105_CANDIDATE |
| Fix Version/s | 10.4 [ 22408 ] |
| Link |
This issue blocks |
| Link |
This issue relates to |
| Link |
This issue blocks |
| Status | In Progress [ 3 ] | Stalled [ 10000 ] |
| Fix Version/s | 10.5 [ 23123 ] | |
| Fix Version/s | 10.3 [ 22126 ] | |
| Fix Version/s | 10.4 [ 22408 ] | |
| Assignee | Thirunarayanan Balathandayuthapani [ thiru ] | Vladislav Vaintroub [ wlad ] |
| Priority | Major [ 3 ] | Critical [ 2 ] |
| Fix Version/s | 10.5 [ 23123 ] |
| Fix Version/s | 10.5 [ 23123 ] |
| Status | Stalled [ 10000 ] | In Progress [ 3 ] |
| Status | In Progress [ 3 ] | Stalled [ 10000 ] |
| Assignee | Vladislav Vaintroub [ wlad ] | Marko Mäkelä [ marko ] |
| Status | Stalled [ 10000 ] | In Progress [ 3 ] |
| Link |
This issue blocks |
| Link |
This issue relates to |
| Link |
This issue relates to |
| Description | [MySQL 8.0.0 split the InnoDB {{buf_pool_t::mutex}}|https://github.com/mysql/mysql-server/commit/2bcc00d11f21fe43ba3c0e0f81d3d9cec44c44a0]. MariaDB should do the same. |
[MySQL 8.0.0 split the InnoDB {{buf_pool_t::mutex}}|https://github.com/mysql/mysql-server/commit/2bcc00d11f21fe43ba3c0e0f81d3d9cec44c44a0]. MariaDB should do something similar.
Instead of introducing more mutexes or radically changing the latching rules of various {{buf_pool_t}} and {{buf_block_t}} data members, I think that it is possible to reduce the contention on {{buf_pool.mutex}} by other means: * Move more code to inline functions of {{buf_pool_t}} or {{buf_page_t}}. * Reduce the amount of mutex release/reacquire dance in {{buf0flu.cc}} and {{buf0rea.cc}}. * Avoid repeated calls to {{page_id_t::fold()}} or {{page_id_t::page_id_t()}}; use {{page_id_t}} directly as loop iterator. * Move {{buf_page_t::flush_type}} to {{IORequest}}. * Split {{buf_page_io_complete()}} into separate ‘read completion’ and ‘write completion’ callbacks. * Avoid holding {{buf_pool.mutex}} during {{buf_pool.page_hash}} operations. Consider removing the debug field {{buf_page_t::in_page_hash}}. * Split operations on {{buf_pool.watch[]}} into two parts. The allocation of {{buf_pool.watch[]}} should be only protected by {{buf_pool.mutex}}, and the {{buf_pool.page_hash}} only by the hash bucket latch. |
| Summary | Split buf_pool_t::mutex | Reduce buf_pool_t::mutex contention |
| Assignee | Marko Mäkelä [ marko ] | Eugene Kosov [ kevg ] |
| Status | In Progress [ 3 ] | In Review [ 10002 ] |
| Link |
This issue relates to |
| Assignee | Eugene Kosov [ kevg ] | Marko Mäkelä [ marko ] |
| Status | In Review [ 10002 ] | Stalled [ 10000 ] |
| Status | Stalled [ 10000 ] | In Progress [ 3 ] |
| Link |
This issue is duplicated by |
| issue.field.resolutiondate | 2020-06-05 09:40:41.0 | 2020-06-05 09:40:41.558 |
| Fix Version/s | 10.5.4 [ 24264 ] | |
| Fix Version/s | 10.5 [ 23123 ] | |
| Resolution | Fixed [ 1 ] | |
| Status | In Progress [ 3 ] | Closed [ 6 ] |
| Link |
This issue causes |
| Link |
This issue relates to |
| Link |
This issue relates to |
| Link |
This issue relates to |
| Link |
This issue causes |
| Link |
This issue causes |
| Link |
This issue causes |
| Link |
This issue relates to |
| Link |
This issue relates to |
| Link |
This issue causes |
| Link |
This issue causes |
| Link |
This issue causes |
| Link |
This issue causes |
| Link | This issue relates to MDEV-24090 [ MDEV-24090 ] |
| Link |
This issue causes |
| Workflow | MariaDB v3 [ 85114 ] | MariaDB v4 [ 133446 ] |
| Link |
This issue causes |
| Link |
This issue causes |
| Link |
This issue causes |
| Link |
This issue relates to |