@Neo: That are definitely good points. Regarding performance, I do not need very high performance for my zfs servers, so I have no comparison to other servers using btrfs/ext4.
What's regarding the quote
it has no real maintenance behind it either any more, as I see it the zfs on linux code seems to be in development steadily. The names of the top contributors Brian Behlendorf or Matthew Ahrens sound familiar to me in terms of respectable international known programmers. So before hearing more about that I assume, the OpenZFS project is very active at the moment.
Code frequency . openzfs/zfs . GitHub
Contributors to openzfs/zfs . GitHub
Update:
Here's a more comprehensive benchmark comparison(from 2015, quite ancient, important because btrfs evolved a lot in recent years) between zfs and btrfs(and partly also with xfs and ext4):
https://www.diva-portal.org/smash/ge...FULLTEXT01.pdf
Quote:
Conclusion
The conclusion of the main findings is that Btrfs showed a high average throughput for most of the
tests, compared to the other file systems. By comparing the results to the results from previous
similar work it shows that the performance has improved greatly in recent years, especially the
multiple disk performance. Btrfs has had big problems with the stability of RAID 5 and RAID 6 and
only in the latest Linux kernel, at the time of writing, was it considered stable. Therefore a RAID 5
setup was also included in the experiments and the results surprisingly showed that Btrfs had
significantly higher average throughput than the other file systems. ZFS however did not perform as
well but the exact reason for this could not be established from the data gathered in the experiments,
but is instead listed as in the future work section (see Section 7.3).
Update-2:
A Performance comparison of zfs vs xfs of Percona - the mysql experts (from 2018)
About ZFS Performance - Percona Database Performance Blog
Quote:
Conclusion
We have seen in this post why the general perception is that ZFS under-performs compared to XFS or EXT4. The presence of B-trees for the files has a big impact on the amount of metadata ZFS needs to handle, especially when the recordsize is small. The metadata consists mostly of the non-leaf pages (or internal nodes) of the B-trees. When properly cached, the performance of ZFS is excellent. ZFS allows you to optimize the use of EBS volumes, both in term of IOPS and size when the instance has fast ephemeral storage devices. Using the ephemeral device of an i3.large instance for the ZFS L2ARC, ZFS outperformed XFS by 66%.
Here another guide on tuning with zfs
I just place it here for others to read. The topic is quite interesting for me right now. The document is a reminder to verify every setting you made in a complex system by testing it and reverting it if it did not improve the situation. He shows a lot of examples which had negative performance impact, like LVM, Compression, Default Proxmox CPU-assignment(kvm64, for performance its better to use host-cpu, but you may sacrifice live-migration capability if you have a mixed hardware pool), Proxmox Storage Driver, ...
ZFS performance tuning - Martin Heiland