Bug #97525 Lack of monitoring visibility when disk is full and slave is paused
Submitted: 6 Nov 2019 22:26 Modified: 7 Nov 2019 13:19
Reporter: Alliumcepa Triplef Email Updates:
Status: Not a Bug Impact on me:
None 
Category:MySQL Server Severity:S3 (Non-critical)
Version:8.0.18 OS:Any
Assigned to: MySQL Verification Team CPU Architecture:Any

[6 Nov 2019 22:26] Alliumcepa Triplef
Description:
When the filessystem run out of space, mysql slave reported both Slave IO and SQL thread running, with seconds behind master being 0. But with Slave_IO_State set to Waiting for disk space.

This means to catch such case in monitoring, one will need to intercept the text string in Slave_IO_State row.

Is it possible to properly catch this case when slave is waiting for disk and paused?

How to repeat:
See description
[7 Nov 2019 13:19] MySQL Verification Team
Hi,

Your monitoring is supposed to detect issues with disk space waaaaaaaaaaaay before this happens so you are really looking at the wrong direction in my opinion. As for the state of the slave, as you found yourself, the Slave_IO_State is giving you the state.