Bug #118888 DDL operation fills the entire disk space and eventually fails
Submitted: 25 Aug 7:19 Modified: 16 Sep 11:29
Reporter: Mehul Todi Email Updates:
Status: Verified Impact on me:
None 
Category:MySQL Server: DDL Severity:S2 (Serious)
Version:8.0.43/8.4.6 OS:Any
Assigned to: CPU Architecture:Any

[25 Aug 7:19] Mehul Todi 
Description:
When executing `CREATE INDEX` with a high value for `--innodb_parallel_read_threads` like 32, the entire disk space is filled up. Once the disk space is exhausted, the command fails and disk usage is back to normal.

How to repeat:
It can be reproduced by running an MTR using:
```
mtr innodb.ddl_large_record --mysqld=--innodb_parallel_read_threads=32
```
[25 Aug 7:21] Mehul Todi 
A possible workaround is to reduce `--innodb_parallel_read_threads` or increase `--innodb_ddl_buffer_size`.
[25 Aug 9:12] MySQL Verification Team 
Thank you for the report and for the test case.
[16 Sep 11:29] Mehul Todi 
Thank you for the report verification. To add some additional details, we want to clarify that this is a regression from 8.0.42/8.4.5 and that we have identified commit - https://github.com/mysql/mysql-server/commit/de289afb0edc89388a6baaed4cdc913cdc891579 to be the root cause for this issue.