modern-fs-benchmark

Good job and clean website interface! Something as exhaustive as this is clearly needed. Some remarks: 1. Why does the CoW button remove XFS from the list? Even https://github.com/fenio/modern-fs-benchmark/blob/main/scrip... mentions it has reflink enabled 2. If you have the time, adding XFS + mdraid + dm-integrity [1] (in bitmap mode) as a comparison point against ZFS RAID-Zx might be an interesting data point. That's what I run, personally. 3. Did you give some thoughts to the I/O scheduler choice? Might matter a lot in some cases. [1] https://www.kernel.org/doc/html/latest/admin-guide/device-ma...

The script[1] shows it's a ZFS mirror (RAID-1) with 8k max record size, as opposed to the default 128k. I say max, as ZFS stores multiples of the physical block size up to the record size. So even with 128k record size, if your file fits in two physical blocks, it will write two physical blocks of data. The `ashift` parameter[2] controls the physical block size, typically 512 bytes or 4k for HDDs. Though higher can be useful on SSDs (but less tested, have seen some bug reports with >4k block sizes). Typically you'd only want to use such small records if you are tuning it for a specific workload, like if you have a database that writes 8k pages. Compression acts on records, so by limiting the record size, you limit the effectiveness of the compression: 1.2 and 1.8 blocks worth of data both gets written as 2 physical blocks. On the upside,

Seems like a bug in benchmark code for that test, opened GH issue https://github.com/fenio/modern-fs-benchmark/issues/35

It's a fair point but it's also possible the person running the tests has a dedicated test hypervisor for this , so that different configurations of filesystems and VMs can be created and destroyed quickly in an automated manner. If it's something as simple as a KVM hypervisor that only runs 1 test VM at a time (with no other load from anything else other than the basic systemd daemons, ssh daemon etc running on the hypervisor), the results could be very close to bare metal. I can see it being very time consuming and annoying to do repeated manual bare metal OS installs and new partitioning/filesystem creation for such a large variety of tests. The author does also say that performance isn't really the main thing but rather, data integrity: https://github.com/fenio/modern-fs-benchmark

What is md-raid10 doing that is so much worse than lvm-raid10? In terms of "I/O" and "responsiveness." It's not really obvious to me from either the linked page or https://github.com/fenio/modern-fs-benchmark . In principle they should be similar?