[MDEV-17025] MariaDB Server 10.1 online ddl abnormal Created: 2018-08-21  Updated: 2018-09-27  Resolved: 2018-09-27

Status: Closed
Project: MariaDB Server
Component/s: Storage Engine - InnoDB
Affects Version/s: 10.1.22
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: wy Assignee: Unassigned
Resolution: Incomplete Votes: 0
Labels: need_feedback
Environment:

MariaDB Server 10.1.22 version in debian 8.2

Attachments: PNG File host-error-file-ddl-info.png     PNG File prometheus-ActiveThreads.png     PNG File prometheus-DDL.png     PNG File prometheus-ProcessStatus.png     PNG File prometheus-SlowQuerys.png    

 Description   

One table do twice DDL change different column name in same one command.

the SQL command like alter table tb1 change column1 column2;alter table tb1 change column3 column4.the syntax without of data type's comment.

column1 is int data type and column3 is varchar(64).

first DDL SQL execute time from 11:07:45 to 11:28:16，but not logging in host error file and severely disrupted production，please see host-error-file-ddl-info.pnp file.
second DDL SQL execute time from 11:28:16 to 11:48:30.

 Comments   
Comment by Elena Stepanova [ 2018-08-29 ]

Sorry, I don't understand the nature of the problem.
Does the ALTER TABLE fail?
Why should it write anything to the error log?
What disrupts production?
Please clarify what the issue is, which result you are getting, why it is incorrect, and what you expect to get instead.

For host-error-file-ddl-info.png, I don't see anything abnormal in it, except for InnoDB being excessively verbose, which is a known issue in 10.1, but in your case with messages only written every few minutes, certainly not critical.

Generated at Thu Feb 08 08:33:21 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.