Annotation of wikisrc/zfs.mdwn, revision 1.19
1.1 gdt 1: # ZFS on NetBSD
2:
1.11 gdt 3: This page attempts to do two things: provide enough orientation and
1.13 wiz 4: pointers to standard ZFS documentation for NetBSD users who are new to
1.11 gdt 5: ZFS, and to describe NetBSD-specific ZFS information. It is
1.1 gdt 6: emphatically not a tutorial or an introduction to ZFS.
7:
8: Many things are marked with \todo because they need a better
9: explanation, and some have question marks, indicating that the
10: statement needs verification.
11:
12: # Documentation Pointers
13:
14: See the man pages for zfs(8) and zpool(8).
15:
1.4 gdt 16: - [Oracle ZFS Administration Manual](https://docs.oracle.com/cd/E26505_01/html/E37384/index.html)
17: - [FreeBSD Handbook ZFS Chapter](https://www.freebsd.org/doc/handbook/zfs.html)
1.14 gdt 18: - [OpenZFS admin docs index page](https://github.com/openzfs/zfs/wiki/Admin-Documentation)
1.1 gdt 19:
20: # Status of ZFS in NetBSD
21:
1.12 gdt 22: ## Sources
23:
24: \todo Verify/fix.
25:
26: Currently, there are multiple ZFS projects/codebases:
27:
28: - ZFS as released under the CDDL (common ancestor)
29: - [OpenZFS](http://www.open-zfs.org/wiki/Main_Page)
30: - [zfsonlinux](https://zfsonlinux.org/)
31: - proprietary ZFS in Solaris (not relevant in open source)
32:
33: OpenZFS is a coordinating project to align open ZFS codebases. There
34: is a notion of a shared core codebase and OS-specific adaptation code.
35:
1.16 gdt 36: See [FreeBSD's history](https://wiki.freebsd.org/ZFSTuningGuide).
1.14 gdt 37:
1.15 gdt 38: \todo Explain the relationship between OpenZFS and zfsonlinux.
39:
1.12 gdt 40: ## NetBSD code history
41:
42: \todo This section really needs help.
43:
44: Before NetBSD 8, NetBSD imported ZFS code from ?
45:
46: Before NetBSD 9, NetBSD imported updated ZFS code from FreeBSD. That
47: FreeBSD code came from ?
1.1 gdt 48:
49: ## NetBSD 8 and earlier
50:
51: While there is some ZFS code, it is old, and seems to have significant
52: problems. If one wants to use ZFS, first upgrade to NetBSD 9. It is
53: unlikely that anyone is interested in helping, other than telling you
54: to upgrade to 9.
55:
1.13 wiz 56: (Reports of how well NetBSD 8 works are welcome on netbsd-users, if it
1.1 gdt 57: can actually be recommended for use.)
58:
59: ## NetBSD 9
60:
61: There have been fixes since 9.0 RELEASE. It is best to upgrade along
1.12 gdt 62: the netbsd-9 branch, but the release should be ok. Most aspects work solidly.
1.1 gdt 63:
1.12 gdt 64: \todo Explain this in terms of versions of FreeBSD OpenZFS and/or
65: zfsonlinux.
1.1 gdt 66:
67: \todo This supports pool version 28/5000 (really true?). Of the
68: feature flags found in modern OpenZFS, \todo are supported.
69:
1.7 gdt 70: Generally, fixes to ZFS in current will be pulled up to 9, but new
1.5 gdt 71: features typically will not be.
72:
1.1 gdt 73: ## NetBSD current
74:
1.5 gdt 75: The ZFS code in current is very similar to that in 9.
76:
1.9 gdt 77: There is initial support for [[ZFS root|wiki/RootOnZFS]], via booting from
1.5 gdt 78: ffs and pivoting.
1.1 gdt 79:
1.14 gdt 80: One can make a ccd using a zvol as a component. See the zvol section below.
1.1 gdt 81:
1.16 gdt 82: ## Things that aren't supported yet
1.1 gdt 83:
1.8 gdt 84: \todo hotswap (maybe - not clear exactly what this means)
1.1 gdt 85:
1.13 wiz 86: \todo direct boot into zfs root (via boot blocks reading zfs)
1.1 gdt 87:
88: ## Architectures
89:
90: Most people seem to be using amd64.
91:
92: To build zfs, one puts MKZFS=yes in mk.conf. This is default on amd64
93: and aarch64 on netbsd-9. In current, it is also default on sparc64.
94:
95: More or less, zfs can be enabled on an architecture when it is known
96: to build and run reliably. (Of course, users are welcome to build it
97: and report.)
98:
99: # NetBSD-specific information
100:
101: ## rc.conf
102:
103: The main configuration is to put zfs=YES in rc.conf, so that the rc.d
1.13 wiz 104: scripts bring up ZFS and mount ZFS file systems.
1.1 gdt 105:
1.14 gdt 106: ## pool locations
107:
108: One can add disks or parts of disks into pools. Methods of specifying
109: areas to be included include:
110:
111: - entire disks (e.g., /dev/rwd0d on amd64)
112: - disklabel partitions (e.g., /dev/sd0e)
113: - wedges (e.g., /dev/dk0)
114:
1.6 gdt 115: ## legacy vs ? mount points
116:
117: \todo Explain, if this is NetBSD specific. Explain consequences, as
118: this seems to have something to do with mount ordering.
119:
1.1 gdt 120: ## mount order
121:
1.13 wiz 122: NetBSD 9 mounts other file systems and then ZFS file systems. This can
1.3 gdt 123: be a problem if /usr/pkgsrc is on ZFS and /usr/pkgsrc/distfiles is on
124: NFS. A workaround is to use noauto and do the mounts in
125: /etc/rc.local.
126:
127: NetBSD current after 20200301 mounts ZFS first. \todo Explain
128: consequences.
1.1 gdt 129:
1.19 ! gdt 130: ## NFS
! 131:
! 132: \todo Verify if this is accurate.
! 133:
! 134: zfs filesystems can be exported over NFS. While there are zfs
! 135: commands that appear to be about controlling exports, they simply
! 136: print information that can be added to exports(5).
! 137:
1.14 gdt 138: ## zvol
139:
140: Within a ZFS pool, the standard approach is to have file systems, but
141: one can also create a zvol, which is a block device of a certain size.
142:
143: \todo The zvol will appear as /dev/???? and can be used in many
144: respects like a slice. However, the system will not read disklabels
145: and gpt labels from a zvol; in this respect it is more like a disklabel
146: partition or wedge than a disk drive.
147:
148: \todo Explain that one can export a zvol via iscsi.
149:
150: \todo Explain if one can swap on a zvol.
151:
152: \todo Explain that one can use ccd to create a normal-looking disk
153: from a zvol. This allows reading a GPT label from the zvol, which is
154: useful in case the zvol had been exported via iscsi and some other
155: system created a label.
156:
1.1 gdt 157: ## TRIM
158:
159: There is some notion of TRIM and zfs using it.
160:
161: \todo Explain how this relates to NetBSD.
162:
163: # Memory usage
164:
165: Basically, ZFS uses lots of memory and most people run it on systems
166: with large amounts of memory. NetBSD works well on systems with
167: comparatively small amounts of memory. So a natural question is how
168: well ZFS works on one's VAX with 2M of RAM :-)
169:
1.11 gdt 170: More seriously, one might ask if it is reasonable to run ZFS on a RPI3
1.1 gdt 171: with 1G of RAM, or even if it is reasonable on a system with 4G.
172:
173: \todo Give ballpark level for minimum sane RAM, and the amount which
174: is cleanly enough.
175:
1.10 gdt 176: For now, a good guess is that a 4G system with only 1T of disk is
1.18 gdt 177: probably ok, and that 1G is very likely not ok.
1.10 gdt 178:
1.2 gdt 179: FreeBSD has some documentation about memory use. There is a notion of
1.18 gdt 180: a minimum of 1G (used for ZFS), and using 1G for 1T of storage, and
181: more if deduplication is enabled. FreeBSD considers all i386 systems
182: to be low memory; this appears to be a clue.
1.1 gdt 183:
1.2 gdt 184: \todo Explain if the FreeBSD sysctl list applies, or if not what we
185: should do instead.
186:
1.7 gdt 187: - [FreeBSD low memory documentation](https://www.freebsd.org/doc/handbook/zfs-advanced.html)
1.1 gdt 188:
189: # Interoperability with other systems
190:
191: \todo Explain pool version and feature flags relationship to FreeBSD,
192: Linux, OpenIndiana/Illumos/?, and ?
193:
1.18 gdt 194: \todo Explain how to configure a pool in terms of version/features for
195: use with particula other systems.
196:
1.14 gdt 197: # Quick Start
198:
199: See the [FreeBSD Quickstart
200: Guide](https://www.freebsd.org/doc/handbook/zfs-quickstart.html); only
201: the first item is NetBSD specific.
202:
203: - Put zfs=YES in rc.conf.
204:
205: - Create a pool as "zpool create pool1 /dev/dk0".
206:
207: - df and see /pool1
208:
1.17 gdt 209: - Really, read the FreeBSD docs and the other linked documentation above.
CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb