Aurora MySQL 3.07.1: incorrect value of the variable Threads_running (18446744073709547770) in global status when using Local read replica write forwarding

0

A similar problem has allegedly already been solved here:

https://docs.aws.amazon.com/AmazonRDS/latest/AuroraMySQLReleaseNotes/AuroraMySQL.Updates.2121.html

"Fixed an issue which can produce an incorrect value of the variable Threads_running in the information_schema and performance_schema global status tables when using write forwarding."

And here

https://docs.aws.amazon.com/AmazonRDS/latest/AuroraMySQLReleaseNotes/AuroraMySQL.Updates.3043.html

"Fixed an issue that provided an incorrect value for the threads_running status variable when using Aurora Global Database."

================================================

Details:

Engine version: 8.0.mysql_aurora.3.07.1

Capacity type: Provisioned

Local read replica write forwarding: Enabled But the sql query (DDL) was executed via Writer.

Our APP ERROR:

SEVERE [liquibase.changelog] ChangeSet ... encountered an exception.
liquibase.exception.UnexpectedLiquibaseException: Percona exited with 29
	at liquibase.ext.percona.PTOnlineSchemaChangeStatement.execute(PTOnlineSchemaChangeStatement.java:251)

[2024-10-23 05:24:45] INFO [liquibase.ext] (in cleanup) 2024-10-23T05:24:45 Error copying rows from `n531`.`PartnerProductAccess` to `n531`.`_PartnerProductAccess_new`: 

Threads_running=18446744073709547770 exceeds its critical threshold 50

================================================

Workaround: manually reboot Aurora instance

After reboot Threads_running = 2

And DDL sql queries was executed successfully

已提問 3 個月前檢視次數 487 次
2 個答案
0

The issue you're experiencing with an incorrect value for the Threads_running variable in Aurora MySQL 3.07.1 when using Local read replica write forwarding is a known problem that has been addressed in previous Aurora MySQL updates.

This specific issue has been fixed in earlier versions of Aurora MySQL, including version 3.06.1 and 3.04.3. In these updates, Amazon addressed the problem of incorrect values for the threads_running status variable when using Aurora Global Database.

The extremely large value you're seeing for Threads_running (18446744073709547770) is likely due to an integer overflow or similar data representation issue, rather than an actual count of running threads.

While the fix was implemented for Aurora Global Database scenarios, it appears that a similar issue may still be present in your specific configuration using Local read replica write forwarding in version 3.07.1.

Your workaround of manually rebooting the Aurora instance, which results in Threads_running returning to a normal value (2 in your case), is a temporary solution. However, this is not ideal for production environments as it requires downtime and manual intervention.

Given that this issue has been addressed in previous versions, it's likely that AWS is aware of the problem and may be working on a fix for version 3.07.1 or future releases. In the meantime, you may want to consider the following actions:

  1. Contact AWS Support to report this specific issue with Aurora MySQL 3.07.1 and Local read replica write forwarding. They may be able to provide a more immediate solution or workaround.

  2. If possible, consider temporarily disabling Local read replica write forwarding to see if it resolves the issue without requiring instance reboots.

  3. Monitor for any new Aurora MySQL engine updates that might address this issue.

  4. If the problem significantly impacts your operations, you might want to consider downgrading to a version where this issue has been confirmed as fixed, such as 3.06.1 or 3.04.3, until a fix for 3.07.1 is available.

Remember to thoroughly test any changes in a non-production environment before applying them to your production systems.
Sources
Aurora MySQL database engine updates 2024-06-26 (version 3.06.1, compatible with MySQL 8.0.34) - Amazon Aurora
Aurora MySQL database engine updates 2024-06-26 (version 3.04.3, compatible with MySQL 8.0.28) - Amazon Aurora

profile picture
已回答 3 個月前
0

Thanks for the artificial intelligence ML Algorithm )), but can anyone report a lost bug fix?

已回答 3 個月前

您尚未登入。 登入 去張貼答案。

一個好的回答可以清楚地回答問題並提供建設性的意見回饋,同時有助於提問者的專業成長。

回答問題指南