Fedora 37 - Performance Issues with MySQL After Upgrading to Systemd 252
After trying multiple solutions online, I still can't figure this out. After upgrading to Fedora 37, I've started experiencing significant performance degradation with my MySQL database. Queries that used to execute in under 100ms now take several seconds. I've checked the system logs and noticed the following warning in the journal: `MySQL[1234]: InnoDB: Warning: Read page with page number 12345 is not found in the buffer pool.` I also see that the `systemd-analyze blame` command shows MySQL taking more time to start than before, which raises some concerns about potential misconfigurations. Here’s how I initially set up my MySQL service: ```ini [Unit] Description=MySQL Server After=network.target [Service] Type=simple User=mysql ExecStart=/usr/bin/mysqld --defaults-file=/etc/my.cnf [Install] WantedBy=multi-user.target ``` I tried increasing the `Innodb_buffer_pool_size` in my `my.cnf` to 2G, but the performance issues persisted. Additionally, I utilized `strace` to log the system calls MySQL was making, and it appears that there's a lot of disk I/O activity happening during query execution, which was not the case before the upgrade. Here’s the relevant part of my `my.cnf`: ```ini [mysqld] innodb_buffer_pool_size = 2G innodb_log_file_size = 256M innodb_flush_log_at_trx_commit = 2 ``` Finally, I've checked for SELinux denials using `audit2allow`, but there are no relevant denials reported. Any insights on how to identify or resolve the performance issues? What steps can I take to diagnose this further or revert to the previous performance level? My team is using Sql for this application. Am I approaching this the right way? Am I approaching this the right way?