Details
-
Bug
-
Status: Closed (View Workflow)
-
Critical
-
Resolution: Incomplete
-
10.4.26
-
OS : Rocky Linux release 8.6 (Green Obsidian), 4.18.0-372.9.1.el8.x86_64
CPU : 2 x 48 AMD EPYC 7643 48-Core Processor
Memory : 1TB
Disk : 7.3TB ssd
Description
I have a super production DB server as above environment.
DB servers consist of 1 master DB server, 2 replication servers.
And there are 3 WAS server and 1 RP server.
Generally, DB server used 10~30% cpu rate.(active db sessions are 30~40)
But DB server is sometimes very slow.( 1~3 times a week)
So actvive db session increase to 3000(max), and cpu rate up to 30%~60%.
And our service have been stopped.
This problem occur at non-peak time.
This is about master DB server.
I've googling for a month, but I didn't get any reason and solution unfortunately.
I doubt innodb_log_buffer_size, but I'm not sure.
Currently, our only solution is that WAS servers httpd deamon stop, and restart after active DB session is 0 from WAS servers.
I attach mariadb variables and innodb status of problem's initial time and innodb status of normal.
What can I check?
I hope this nightmare is ended ASAP.
Thanks in advance.
Attachments
Issue Links
- duplicates
-
MDEV-31044 MariaDB(10.4) is sometimes very slow on Centos
- Closed