Bug #91321 FR: status var to show innodb io capacity target

• Submitted: 20 Jun 2018 0:28
• Modified: 21 Jun 2018 4:16
• Reporter: Morgan Tocker
• Status: Verified
• Category: MySQL Server: InnoDB storage engine
• Severity: S4 (Feature request)
• Version: 8.0.11
• OS: Any
• CPU Architecture: Any

[20 Jun 2018 0:28] Morgan Tocker

Description:
(Background context) By default InnoDB will use innodb_io_capacity IOs for page cleaner threads as part of background operations.

If the redo log is constrained (i.e. innodb_adaptive_flushing_lwm is reached and innodb_adaptive_flushing=1) or the buffer pool is dirty (innodb_max_dirty_pages_pct_lwm is reached) it will increase the io target up to innodb_io_capacity_max (defaults to 10x io-capacity).

So my understanding is that effective io-capacity target is between innodb_io_capacity and innodb_io_capacity_max.  It would be useful to DBAs to export this target number to rationalize server activity.

How to repeat:
As a simple tuning point: A DBA might consider increasing the redo log size if they see the target come close to the innodb_io_capacity_max; or increase innodb_io_capacity if the base number is unrealistic.

[21 Jun 2018 4:16] MySQL Verification Team

Hello Morgan,

Thank you for the feature request!

Regards,
Umesh