[MDEV-20915] Increased waits for synch/mutex/aria/PAGECACHE::cache_lock Created: 2019-10-29  Updated: 2019-11-13  Resolved: 2019-11-13

Status: Closed
Project: MariaDB Server
Component/s: Storage Engine - Aria
Affects Version/s: 10.2.21
Fix Version/s: N/A

Type: Bug Priority: Major
Reporter: Nicolae Popa Assignee: Axel Schwenke
Resolution: Not a Bug Votes: 0
Labels: None
Environment:

AWS RDS db.m5.4xlarge
16vCPU 64GB RAM


 Description   

About a couple of days ago, a query that was normally working fast started crashing and getting stuck in "Sending data". These eventually pile up and freeze the db.
The query in question is a select on a more complicated view that contains some joins and a union.
I have already created a thread on aws forums about the issue that I'm seeing and added more information as well as performance insights.
https://forums.aws.amazon.com/thread.jspa?messageID=921037#921037

 Comments   
Comment by Nicolae Popa [ 2019-11-13 ]

Feel free to close this bug. After long investigations I discovered that the max_tmp_tables was reached spilling the tables to disk hence the aria pagecache wait increase. It was a miss-configuration on my side and not aria itself

Comment by Axel Schwenke [ 2019-11-13 ]

problem cause by misconfiguration - tmp tables being materialized to disk instead of being held in memory

Generated at Thu Feb 08 09:03:11 UTC 2024 using Jira 8.20.16#820016-sha1:9d11dbea5f4be3d4cc21f03a88dd11d8c8687422.