Open thatguyatgithub opened 1 year ago
You may find the notes in #10253 germane to your interests.
What have you done to your pool to make 300GB scrub run for a month? Are you using some incredibly small blocks and having terrible fragmentation? Of your pool is constantly busy with workload? Or something is failing?
Hello. I'm using zfs-2.1.6-2 on Debian with Linux 6.0.0-2-amd64 system.
When attempting to speed up the scrub process, I was not able to find any of the referred zfs_scrub_delay and zfs_scan_idle parameters on the sysfs at /sys/module/zfs/parameters/*.
I'm kind of desperate because the scrub process is increasingly taking more and more, and the 300G of data will take longer than a month by this rate :-(
I was following this https://openzfs.github.io/openzfs-docs/Performance%20and%20Tuning/ZIO%20Scheduler.html, which appears to be the latest reference. Any idea which parameters have superseded them?