Bug #71382 Please provide graphs to capture P_S_*_lost values
Submitted: 14 Jan 2014 22:39 Modified: 16 Jan 2014 18:53
Reporter: Simon Mudd (OCA) Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Enterprise Monitor: Server Severity:S4 (Feature request)
Version:3.0.5 OS:Any
Assigned to: CPU Architecture:Any

[14 Jan 2014 22:39] Simon Mudd
Description:
MySQL 5.6 provides several counters on data lost by P_S. E.g.

root@myhost [performance_schema]> show global status like 'performance_schema%';
+-----------------------------------------------+---------+
| Variable_name                                 | Value   |
+-----------------------------------------------+---------+
| Performance_schema_accounts_lost              | 2232186 |
| Performance_schema_cond_classes_lost          | 0       |
| Performance_schema_cond_instances_lost        | 0       |
| Performance_schema_digest_lost                | 0       |
| Performance_schema_file_classes_lost          | 0       |
| Performance_schema_file_handles_lost          | 0       |
| Performance_schema_file_instances_lost        | 0       |
| Performance_schema_hosts_lost                 | 476719  |
| Performance_schema_locker_lost                | 0       |
| Performance_schema_mutex_classes_lost         | 0       |
| Performance_schema_mutex_instances_lost       | 0       |
| Performance_schema_rwlock_classes_lost        | 0       |
| Performance_schema_rwlock_instances_lost      | 0       |
| Performance_schema_session_connect_attrs_lost | 0       |
| Performance_schema_socket_classes_lost        | 0       |
| Performance_schema_socket_instances_lost      | 0       |
| Performance_schema_stage_classes_lost         | 0       |
| Performance_schema_statement_classes_lost     | 0       |
| Performance_schema_table_handles_lost         | 0       |
| Performance_schema_table_instances_lost       | 0       |
| Performance_schema_thread_classes_lost        | 0       |
| Performance_schema_thread_instances_lost      | 0       |
| Performance_schema_users_lost                 | 0       |
+-----------------------------------------------+---------+
23 rows in set (0.00 sec)

These values should be at 0. If they're not it indicates a P_S sizing "misconfiguration" that's probably not intended.

How to repeat:
See above.

Suggested fix:
Make a graph in MEM which reveals these "lost" events so we can see if any of them is > 0 and then identify which configuration adjustment may be needed.

for one of these the host_cache_size is dynamic so easy to fix.

Advisors for this might also be useful, especially if a good pointer to which values would need to be adjusted where.  The graphs are good as even if the advisors are not enabled you can "see" any issues immediately.
[16 Jan 2014 18:53] MySQL Verification Team
This is a justified feature request.