[Bug 261212] Update the ZFS chapter (20) of the FreeBSD Handbook, and other OpenZFS-related pages

From: <bugzilla-noreply_at_freebsd.org>
Date: Wed, 02 Feb 2022 23:29:46 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=261212

Graham Perrin <grahamperrin@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Update the ZFS chapter (20) |Update the ZFS chapter (20)
                   |of the FreeBSD Handbook     |of the FreeBSD Handbook,
                   |                            |and other OpenZFS-related
                   |                            |pages

--- Comment #1 from Graham Perrin <grahamperrin@gmail.com> ---
Not the FreeBSD Handbook alone … 


Manual pages
============

<https://www.freebsd.org/cgi/man.cgi?query=zfs&sektion=4&manpath=FreeBSD> 

* finds nothing for zfs(4) in FreeBSD 13.0. 

Compare with OpenZFS <https://openzfs.github.io/openzfs-docs/man/4/zfs.4.html>: 

> zfs — tuning of the ZFS kernel module

For zfs(4) in ports
<https://www.freebsd.org/cgi/man.cgi?query=zfs&sektion=4&manpath=FreeBSD-Ports>
– more specifically,
<https://www.freebsd.org/cgi/man.cgi?query=zfs&sektion=4&manpath=FreeBSD+Ports+13.0>
in 'FreeBSD Ports 13.0' – we have, for example: 

> zfs_arc_max=0B (ulong)

There's a corresponding entry in the OpenZFS page for zfs(4). 


Actual tunables
===============

FreeBSD 13.0-RELEASE includes this variable: 

    vfs.zfs.arc.max

– the value of which defaults to: 

    0

The name of the variable in FreeBSD is comparable to the name of the
parameter/variable in original OpenZFS documentation. Fair enough. 


FreeBSD Handbook
================

<https://docs.freebsd.org/en/books/handbook/book/#zfs-advanced-tuning-arc_max>
describes this variable, with a different name, which probably predates OpenZFS
in FreeBSD: 

    vfs.zfs.arc_max

No mention of    0    and there's this: 

> …  all RAM but 1 GB, or 5/8 of all RAM, whichever is more. 
> Use a lower value if …

The word "but" is not ideal ("minus" would be better); and the reader might
wonder: 

   Lower than what?

-- 
You are receiving this mail because:
You are the assignee for the bug.