Details
-
Bug
-
Status: Closed (View Workflow)
-
Major
-
Resolution: Fixed
-
10.11.3, 10.6.13, 10.8.8, 10.9.6, 10.10.4
Description
In MDEV-26827, some monitoring was changed.
Some status variables were changed to point directly to the underlying variables, like this:
{"buffer_pool_read_requests", &buf_pool.stat.n_page_gets, SHOW_SIZE_T}, |
Unlike all other variables in buf_pool.stat this particular one is divided into shards for performance reasons (MDEV-21212). So, the status variable would only reflect the contents of the first shard and not the entire counter. Thanks to dragle for noticing this.
Note: This counter is inaccurate by design. If we protected it correctly, it would introduce a performance bottleneck.
Attachments
Issue Links
- is caused by
-
-
- Closed
-
- relates to
-
-
- Closed
-
Activity
| Field | Original Value | New Value |
|---|---|---|
| Link |
This issue is caused by |
| issue.field.resolutiondate | 2023-05-19 13:18:03.0 | 2023-05-19 13:18:03.65 |
| Fix Version/s | 10.6.14 [ 28914 ] | |
| Fix Version/s | 10.9.7 [ 28916 ] | |
| Fix Version/s | 10.10.5 [ 28917 ] | |
| Fix Version/s | 10.11.4 [ 28918 ] | |
| Fix Version/s | 11.0.3 [ 28920 ] | |
| Fix Version/s | 11.1.2 [ 28921 ] | |
| Fix Version/s | 10.6 [ 24028 ] | |
| Fix Version/s | 10.9 [ 26905 ] | |
| Fix Version/s | 10.10 [ 27530 ] | |
| Fix Version/s | 10.11 [ 27614 ] | |
| Fix Version/s | 11.0 [ 28320 ] | |
| Fix Version/s | 11.1 [ 28549 ] | |
| Resolution | Fixed [ 1 ] | |
| Status | Open [ 1 ] | Closed [ 6 ] |
| Link | This issue blocks MENT-1817 [ MENT-1817 ] |
| Link | This issue blocks MENT-1818 [ MENT-1818 ] |
| Link | This issue blocks MENT-1824 [ MENT-1824 ] |
This bug still appears to exist in 10.9.8.
MariaDB [(none)]> select count from information_schema.INNODB_METRICS where name = 'buffer_pool_read_requests';
-------
| count |
-------
| 83743 |
MariaDB [(none)]> show global status like 'Innodb_buffer_pool_read_requests';
---------------------------------------+
| Variable_name | Value |
---------------------------------------+
| Innodb_buffer_pool_read_requests | 0 |
---------------------------------------+
| Link |
This issue relates to |
rogerd007 note the 10.9 series (and 10.10) are out of maintaince as they where short term support only.
For maintaince policy - see https://mariadb.org/about/#maintenance-policy
Thanks, I believe this is taken care of in the 10.6.14 build you pointed me to in a separate thread. Unfortunately the only machine I had that displayed the problem was reverted early this week to 10.6.12, and when I try to load new 10.6.13 instances in VMs they don't display this behavior.
But not seeing the issue consistently would seem to make sense based on your explanation, and I'm not seeing it in the 10.6.14 build either, so this should be good.