Annotation of wikisrc/zfs.mdwn, revision 1.35

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
1.28      gdt         9: explanation, and some have question marks
1.1       gdt        10: 
                     11: # Documentation Pointers
                     12: 
1.34      gdt        13: See the man pages for zfs(8), zpool(8).  Also see zdb(8), if only for
                     14: seeing pool config info when run with no arguments.
1.1       gdt        15: 
1.27      gdt        16:   - [OpenZFS Documentation](https://openzfs.github.io/openzfs-docs/)
                     17:   - [OpenZFS admin docs index page](https://github.com/openzfs/zfs/wiki/Admin-Documentation)
                     18:   - [FreeBSD Handbook ZFS Chapter](https://www.freebsd.org/doc/handbook/zfs.html)
1.4       gdt        19:   - [Oracle ZFS Administration Manual](https://docs.oracle.com/cd/E26505_01/html/E37384/index.html)
1.24      gdt        20:   - [Wikipedia](https://en.wikipedia.org/wiki/ZFS)
1.21      gdt        21: 
1.1       gdt        22: # Status of ZFS in NetBSD
                     23: 
1.27      gdt        24: ## NetBSD 8
1.23      gdt        25: 
1.27      gdt        26: NetBSD 8 has an old version of ZFS, and it is not recommended for use
                     27: at all.  There is no evidence that anyone is interested in helping
                     28: with ZFS on 8.  Those wishing to use ZFS on NetBSD 8 should therefore
                     29: update to NetBSD 9.
1.1       gdt        30: 
                     31: ## NetBSD 9
                     32: 
1.27      gdt        33: NetBSD-9 has ZFS that is considered to work well.  There have been
                     34: fixes since 9.0_RELEASE.  As always, people running NetBSD 9 are
                     35: likely best served by the most recent version of the netbsd-9 stable
                     36: branch.  As of 2021-02, ZFS in the NetBSD 9.1 release is very close to
                     37: netbsd-9.
1.1       gdt        38: 
1.32      gdt        39: ### Native blocksize
                     40: 
                     41: ZFS attempts to find out the native blocksize for a disk when using it
                     42: in a pool; this is almost always 512 or 4096.  Somewhere between 9.0
                     43: and 9.1, at least some disks on some controllers that used to report
                     44: 512 now report 4096.  This provokes a blocksize mismatch warning.
                     45: 
                     46: Given that the native blocksize of the disk didn't change, and things
                     47: seemed OK using the 512 emulated blocks, the warning is likely not
                     48: critical.  However, it is also likely that rebuilding the pool with
                     49: the 4096 blocksize is likely to result in better behavior because ZFS
                     50: will only try to do 4096-byte writes.  \todo Verify this and find the
                     51: actual change and explain better.
                     52: 
1.27      gdt        53: ## NetBSD-current
1.1       gdt        54: 
1.27      gdt        55: NetBSD-current (as of 2021-02) has similar ZFS code to 9.
1.5       gdt        56: 
1.9       gdt        57: There is initial support for [[ZFS root|wiki/RootOnZFS]], via booting from
1.5       gdt        58: ffs and pivoting.
1.1       gdt        59: 
1.27      gdt        60: ## NetBSD/xen special issues
1.1       gdt        61: 
1.27      gdt        62: In NetBSD-9, MAXPHYS is 64KB in most places, but because of xbd(4) it
                     63: is set to 32KB for XEN kernels.  Thus the standard zfs kernel modules
                     64: do not work under xen.  In NetBSD-current, xbd(4) supports 64 KB
                     65: MAXPHYS and this is no longer an issue.
1.1       gdt        66: 
1.27      gdt        67: Xen and zfs on current are reported to work well together, as of 2021-02.
1.1       gdt        68: 
                     69: ## Architectures
                     70: 
                     71: Most people seem to be using amd64.
                     72: 
                     73: To build zfs, one puts MKZFS=yes in mk.conf.  This is default on amd64
                     74: and aarch64 on netbsd-9.  In current, it is also default on sparc64.
                     75: 
                     76: More or less, zfs can be enabled on an architecture when it is known
                     77: to build and run reliably.  (Of course, users are welcome to build it
                     78: and report.)
                     79: 
1.27      gdt        80: # Quick Start
                     81: 
                     82: See the [FreeBSD Quickstart
                     83: Guide](https://www.freebsd.org/doc/handbook/zfs-quickstart.html); only
                     84: the first item is NetBSD specific.
                     85: 
                     86:   - Put zfs=YES in rc.conf.
                     87: 
                     88:   - Create a pool as "zpool create pool1 /dev/dk0".
                     89: 
                     90:   - df and see /pool1
                     91: 
                     92:   - Create a filesystem mounted on /n0 as "zfs create -o
                     93:     mountpoint=/n0 pool1/n0".
                     94: 
                     95:   - Go back and read the documentation and start over.
                     96: 
1.1       gdt        97: # NetBSD-specific information
                     98: 
                     99: ## rc.conf
                    100: 
                    101: The main configuration is to put zfs=YES in rc.conf, so that the rc.d
1.13      wiz       102: scripts bring up ZFS and mount ZFS file systems.
1.1       gdt       103: 
1.14      gdt       104: ## pool locations
                    105: 
                    106: One can add disks or parts of disks into pools.  Methods of specifying
                    107: areas to be included include:
                    108: 
1.29      gdt       109:   - entire disks (e.g., /dev/wd0d on amd64, or /dev/wd0 which has the same major/minor)
1.14      gdt       110:   - disklabel partitions (e.g., /dev/sd0e)
                    111:   - wedges (e.g., /dev/dk0)
                    112: 
1.33      gdt       113: Information about created or imported pools is stored in
                    114: /etc/zfs/zpool.cache.
                    115: 
                    116: ## pool importing problems
                    117: 
                    118: While one can "zpool pool0 /dev/wd0f" and have a working pool, this
                    119: pool cannot be exported and imported straigthforwardly.  "zpool
1.35    ! gdt       120: export" works fine, and deletes zpool.cache.  "zpool import", however,
        !           121: only looks at entire disks (e.g. /dev/wd0), and might look at slices
        !           122: (e.g. /dev/dk0).  It does not look at partitions like /dev/wd0f, and
        !           123: there is no way on the command line to ask that specific devices be
        !           124: examined.  Thus, export/import fails for pools with disklabel
        !           125: partitions.
1.33      gdt       126: 
                    127: One can make wd0 be a link to wd0f temporarily, and the pool will then
                    128: be importable.  However, "wd0" is stored in zpool.cache and on the
                    129: next boot that will attempt to be used.  This is obviously not a good
                    130: approach.
                    131: 
                    132: One an mkdir e.g. /etc/zfs/pool0 and in it have a symlink to
                    133: /dev/wd0f.  Then, zpool import -d /etc/zfs/pool0 will scan
                    134: /etc/zfs/pool0/wd0f and succeed.  The resulting zpool.cache will have
                    135: that path, but having symlinks in /etc/zfs/POOLNAME seems acceptable.
                    136: 
                    137: \todo Determine a good fix, perhaps man page changes only, fix it
                    138: upstream, in curent, and in 9, before removing this discussion.
                    139: 
1.1       gdt       140: ## mount order
                    141: 
1.13      wiz       142: NetBSD 9 mounts other file systems and then ZFS file systems.  This can
1.3       gdt       143: be a problem if /usr/pkgsrc is on ZFS and /usr/pkgsrc/distfiles is on
                    144: NFS.  A workaround is to use noauto and do the mounts in
                    145: /etc/rc.local.
                    146: 
1.27      gdt       147: NetBSD current after 20200301 mounts ZFS first.  The same issues and
                    148: workarounds apply in different circumstances.
1.1       gdt       149: 
1.19      gdt       150: ## NFS
                    151: 
1.31      gdt       152: zfs filesystems can be exported via NFS, simply by placing them in
                    153: /etc/exports like any other filesystem.
1.27      gdt       154: 
1.31      gdt       155: The "zfs share" command adds a line for each filesystem with the
                    156: sharenfs property set to /etc/zfs/exports, and "zfs unshare" removes
1.33      gdt       157: it.  This file is ignored on NetBSD-9 and current before 20210216; on
                    158: current after 20210216 those filesystems should be exported (assuming
                    159: NFS is enabled).  It does not appear to be possible to set options
                    160: like maproot and network restrictions via this method.
1.31      gdt       161: 
1.33      gdt       162: On current before 20210216, a remote mkdir of a filesystem mounted via
1.31      gdt       163: -maproot=0:10 causes a kernel NULL pointer dereference.  This is now
1.33      gdt       164: fixed.
                    165: 
1.14      gdt       166: ## zvol
                    167: 
                    168: Within a ZFS pool, the standard approach is to have file systems, but
                    169: one can also create a zvol, which is a block device of a certain size.
                    170: 
                    171: \todo The zvol will appear as /dev/???? and can be used in many
                    172: respects like a slice.  However, the system will not read disklabels
                    173: and gpt labels from a zvol; in this respect it is more like a disklabel
                    174: partition or wedge than a disk drive.
                    175: 
                    176: \todo Explain that one can export a zvol via iscsi.
                    177: 
                    178: \todo Explain if one can swap on a zvol.
                    179: 
                    180: \todo Explain that one can use ccd to create a normal-looking disk
                    181: from a zvol.  This allows reading a GPT label from the zvol, which is
                    182: useful in case the zvol had been exported via iscsi and some other
                    183: system created a label.
                    184: 
1.1       gdt       185: # Memory usage
                    186: 
                    187: Basically, ZFS uses lots of memory and most people run it on systems
                    188: with large amounts of memory.  NetBSD works well on systems with
                    189: comparatively small amounts of memory.  So a natural question is how
1.27      gdt       190: well ZFS works on one's VAX with 2M of RAM :-) More seriously, one
                    191: might ask if it is reasonable to run ZFS on a RPI3 with 1G of RAM, or
                    192: if it is reasonable on a system with 4G.
                    193: 
                    194: The prevailing wisdom is more or less that ZFS consumes 1G plus 1G per
                    195: 1T of disk.  32-bit architectures are viewed as too small to run ZFS.
                    196: 
                    197: Besides RAM, zfs requires that architecture kernel stack size is at
                    198: least 12KB or more -- some operations cause stack overflow with 8KB
                    199: kernel stack. On NetBSD, the architectures with 16KB kernel stack are
                    200: amd64, sparc64, powerpc, and experimental ia64, hppa. mac68k and sh3
                    201: have 12KB kernel stack. All others use only 8KB stack, which is not
                    202: enough to run zfs.
                    203: 
                    204: NetBSD has many statistics provided via sysctl; see "sysctl
                    205: kstat.zfs".
                    206: 
                    207: FreeBSD has tunables that NetBSD does not seem to have, described in
                    208: [FreeBSD Handbook ZFS Advanced
                    209: section](https://docs.freebsd.org/en/books/handbook/zfs/#zfs-advanced).
1.1       gdt       210: 
1.27      gdt       211: # Interoperability with other systems
1.1       gdt       212: 
1.27      gdt       213: Modern ZFS uses pool version 5000 and feature flags.
1.1       gdt       214: 
1.27      gdt       215: It is in general possible to export a pool and them import the pool on
                    216: some other system, as long as the other system supports all the used
                    217: features.
1.10      gdt       218: 
1.27      gdt       219: \todo Explain how to do this and what is known to work.
1.26      wiki      220: 
1.27      gdt       221: \todo Explain feature flags relationship to FreeBSD, Linux, iIllumos,
                    222: macOS.
1.1       gdt       223: 
1.27      gdt       224: # Sources of ZFS code
1.2       gdt       225: 
1.27      gdt       226: Currently, there are multiple ZFS projects and codebases:
1.1       gdt       227: 
1.27      gdt       228:   - [OpenZFS](http://www.open-zfs.org/wiki/Main_Page)
1.30      wiz       229:   - [openzfs repository](https://github.com/openzfs/zfs)
1.27      gdt       230:   - [zfsonlinux](https://zfsonlinux.org/)
                    231:   - [OpenZFS on OS X ](https://openzfsonosx.org/) [repo](https://github.com/openzfsonosx)
                    232:   - proprietary ZFS in Solaris (not relevant in open source)
                    233:   - ZFS as released under the CDDL (common ancestor, now of historical interest)
1.1       gdt       234: 
1.27      gdt       235: OpenZFS is a coordinating project to align open ZFS codebases.  There
                    236: is a notion of a shared core codebase and OS-specific adaptation code.
1.1       gdt       237: 
1.27      gdt       238:   - [zfsonlinux relationship to OpenZFS](https://github.com/openzfs/zfs/wiki/OpenZFS-Patches)
1.28      gdt       239:   - FreeBSD more or less imports code from openzfs and pushes back fixes. \todo Verify this.
                    240:   - NetBSD has imported code from FreeBSD.
                    241:   - The status of ZFS on macOS is unclear (2021-02).

CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb