Re: GPT Partition

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


On 2023-09-15 2:07 p.m., Peter Boy wrote:
Am 15.09.2023 um 17:23 schrieb Bill Cunningham <bill.cu1234@xxxxxxxxx>:
WHat is the reason Peter behind xfs being used on the server edition and btrfs on the workstation? I pretty much stick with ext3. I don't even use ext4 really. I've never used xfs.
It is basically about data protection, performances, reliability and easy administration (see So, the Fedora server editions (i.e. Server and CoreOS) use LVM/xfs.   

And please, think twice when you read something like "BTRFS protects us from "silent" corruption of files, which is more of an issue with large volumes of data“ or „... large organizations with many users, btrfs is expected to reduce problems with data corruption…“. It’s more kind of marketing speech than any valid decision criteria or technically based argument. If you use the search engine of your preference you will find a lot of detailed and and technically based discussions of Fedora and Red Hat engineers about the topic. As in most cases, there is no „one absolute truth“ about filesystems as many missionaries claim again and again. It is a question of weighing and criteria for a use case or also for a type of use cases. 

Umm, no.

Fedora server uses LVM because that's what most people have upgraded from.  It adds another data layer to the i/o stack, increasing on-disk complexity.  It is normally required under XFS to provide the missing pieces like RAID levels and snapshots, as XFS was designed in the 70s with hardware RAID in use.  Its also what most sysadmins were trained to use, and its hard to change old habits.  Its solid, but really old tech that BTRFS and ZFS can almost always do better.  Unlike LVM, RAID-5/6 is currently a problem for BTRFS, as it has the write-hole bug that almost all hardware and software implementations also have with the exception of ZFS.  Disks are really cheap, so 2 or 3-way mirroring or RAID-1 are currently the ways to go on BTRFS.  It should be noted that imho the RAID-Z levels on ZFS are superior to all other solutions for reliable data preservation and performance.

Fedora server uses XFS because that's what RHEL and therefore the certified sysadmins use.  Consumer disks are actually more reliable than enterprise disks, but stall for very long periods when re-reading failing sectors.  So, BTRFS actually works better on enterprise disks, as the stall is far smaller.  It is probably the #3 or 4 filesystem around for performance and reliability, but managing it is positively arcane.  There are a number of normal admin operations that are very difficult using XFS, such as shrinking a filesystem (even by a couple of sectors to use a replacement disk).  It also does no runtime error detection/correction of your data, so you depend upon RAID hardware or LVM to do that for you.  If you get an error detected during one of these repair sweeps, recovery is usually no better than an uncorrectable multi-bit BTRFS or ZFS failure.  Putting XFS on a single disk is very questionable as a result.

Fedora desktop uses BTRFS by default for a number of really good reasons.  BTRFS detects bit-rot on the fly.  With mirrored or RAIDed disks it can also correct that bit-rot on the fly.  XFS cannot do that, and requires weekly error detection work.  Putting ZFS or BTRFS on RAID hardware actually makes everything slower, as they do a better, faster and more reliable job in software.  BTRFS and ZFS also have many operational advantages, like much faster migrations, near-instantaneous snapshotting and rollback (LVM takes hours to do the same), and much faster off-machine backups.

While ZFS is the gold standard for reliable filesystems, with the exception of the Ubuntu and Oracle platforms, it cannot be used without paying Oracle lots of money.  BTRFS reimplements much of ZFS in a legally unencumbered codebase.  It is unclear why Fedora has not moved server installs to BTRFS by default, as the advantages in complexity, training and data reliability are huge.  Those all make a BTRFS or ZFS server cheaper to operate, sometimes by a considerable margin.  I know all the RHEL sysadmin people will come out of the woodwork and start shouting about that statement, but just maybe they are incorrect.  YMMV of course.

users mailing list -- users@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to users-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct:
List Guidelines:
List Archives:
Do not reply to spam, report it:
[Index of Archives]     [Older Fedora Users]     [Fedora Announce]     [Fedora Package Announce]     [EPEL Announce]     [EPEL Devel]     [Fedora Magazine]     [Fedora Summer Coding]     [Fedora Laptop]     [Fedora Cloud]     [Fedora Advisory Board]     [Fedora Education]     [Fedora Security]     [Fedora Scitech]     [Fedora Robotics]     [Fedora Infrastructure]     [Fedora Websites]     [Anaconda Devel]     [Fedora Devel Java]     [Fedora Desktop]     [Fedora Fonts]     [Fedora Marketing]     [Fedora Management Tools]     [Fedora Mentors]     [Fedora Package Review]     [Fedora R Devel]     [Fedora PHP Devel]     [Kickstart]     [Fedora Music]     [Fedora Packaging]     [Fedora SELinux]     [Fedora Legal]     [Fedora Kernel]     [Fedora OCaml]     [Coolkey]     [Virtualization Tools]     [ET Management Tools]     [Yum Users]     [Yosemite News]     [Gnome Users]     [KDE Users]     [Fedora Art]     [Fedora Docs]     [Fedora Sparc]     [Libvirt Users]     [Fedora ARM]

  Powered by Linux