Performance tuning for GFS2 filesystems on hosts with large memory
Issue
- My cluster has large spikes in I/O activity when it comes to flushing data back to storage.
- GFS2 journal log flushing causes long pauses and or stalls for my application’s ability to write to filesystem as dirty data is written back to the underlying disk.
- How can I change the default values for
vm.dirty_ratioandvm.dirty_background_ratioin the page cache to smooth out page cache flushing during heavy writing to the filesystem?
Environment
- Red Hat Enterprise Linux 7 (with the High Availability Add-on)
- Red Hat Enterprise Linux 8 (with the High Availability Add-on)
- Red Hat Enterprise Linux 9 (with the High Availability Add-on)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.
