btrfs-progs: docs: enhance btrfs-filesystem manual page

Wording, formatting, explanations, examples.

Signed-off-by: David Sterba <dsterba@suse.com>

1 files changed, 175 insertions, 39 deletions

diff --git a/Documentation/btrfs-filesystem.asciidoc b/Documentation/btrfs-filesystem.asciidoc
index 600bbac0..26126175 100644
--- a/Documentation/btrfs-filesystem.asciidoc
+++ b/Documentation/btrfs-filesystem.asciidoc
@@ -3,7 +3,7 @@ btrfs-filesystem(8)
 
 NAME
 ----
-btrfs-filesystem - control btrfs filesystem
+btrfs-filesystem - command group of btrfs that usually work on the whole filesystem
 
 SYNOPSIS
 --------
@@ -11,14 +11,49 @@ SYNOPSIS
 
 DESCRIPTION
 -----------
-*btrfs filesystem* is used to do the filesystem level control jobs, including
-all the regular filesystem operations like setting/getting label,
-resizing, defragment.
+*btrfs filesystem* is used to do the whole filesystem level tasks, including
+all the regular filesystem operations like resizing, space stats, label
+setting/getting, and defragmentation.
 
 SUBCOMMAND
 ----------
 *df* [options] <path>::
-Show space usage information for a mount point.
+Show a terse summary information about allocation of block group types of a given
+mount point. The original purpose of this command was a debugging helper. The
+output needs to be further interpreted and is not suitable for quick overview.
++
+--
+An example with description:
+
+* device size: '1.9TiB', one device, no RAID
+* filesystem size: '1.9TiB'
+* created with: 'mkfs.btrfs -d single -m single'
+--
++
+------------------------------
+$ btrfs filesystem df /path
+Data, single: total=1.15TiB, used=1.13TiB
+System, single: total=32.00MiB, used=144.00KiB
+Metadata, single: total=12.00GiB, used=6.45GiB
+GlobalReserve, single: total=512.00MiB, used=0.00B
+------------------------------
++
+--
+* 'Data', 'System' and 'Metadata' are separeate block group types.
+'GlobalReserve' is an artificial and internal emergency space, see below.
+* 'single' -- the allocation profile, defined at mkfs time
+* 'total' -- sum of space reserved for
+all allocation profiles of the given type, ie. all Data/single. Note that it's
+not total size of filesystem.
+* 'used' -- sum of used space of the above, ie. file extents, metadata blocks
+--
++
+'GlobalReserve' is an artificial and internal emergency space. It is used eg.
+when the filesystem is full. Its 'total' size is dynamic based on the
+filesystem size, usually not larger than 512MiB, 'used' may fluctuate.
++
+The global block reserve is accounted within Metadata. In case the filesystem
+metadata are exhausted, 'GlobalReserve/total + Metadata/used = Metadata/total'.
 +
 `Options`
 +
@@ -44,7 +79,7 @@ show sizes in TiB, or TB with --si
 If conflicting options are passed, the last one takes precedence.
 
 *defragment* [options] <file>|<dir> [<file>|<dir>...]::
-Defragment file data and/or directory metadata *online*.
+Defragment file data on a mounted filesytem.
 +
 If '-r' is passed, files in dir will be defragmented recursively.
 The start position and the number of bytes to defragment can be specified by
@@ -57,7 +92,7 @@ You can also turn on compression in defragment operations.
 +
 WARNING: Defragmenting with Linux kernel versions < 3.9 or ≥ 3.14-rc2 as well as
 with Linux stable kernel versions ≥ 3.10.31, ≥ 3.12.12 or ≥ 3.13.4 will break up
-the ref-links of CoW data (for example files copied with `cp --reflink`,
+the ref-links of COW data (for example files copied with `cp --reflink`,
 snapshots or de-duplicated data).
 This may cause considerable increase of space usage depending on the broken up
 ref-links.
@@ -65,11 +100,11 @@ ref-links.
 `Options`
 +
 -v::::
-be verbose
+be verbose, print file names as they're submitted for defragmentation
 -c[<algo>]::::
 compress file contents while defragmenting. Optional argument selects the compression
 algorithm, 'zlib' (default) or 'lzo'. Currently it's not possible to select no
-compression.
+compression. See also section 'EXAMPLES'.
 -r::::
 defragment files recursively in given directories
 -f::::
@@ -77,45 +112,51 @@ flush data for each file before going to the next file. This will limit the amou
 of dirty data to current file, otherwise the amount cumulates from several files
 and may increase system load.
 -s <start>[kKmMgGtTpPeE]::::
-defragment only from byte <start> onward
+defragmentation will start from the given offset, default is beginning of a file
 -l <len>[kKmMgGtTpPeE]::::
-defragment only up to <len> bytes
+defragment only up to 'len' bytes, default is the file size
 -t <size>[kKmMgGtTpPeE]::::
-target extent size, do not touch extents bigger than <size>
+target extent size, do not touch extents bigger than 'size'
 +
-For <start>, <len>, <size> it is possible to append
+For 'start', 'len', 'size' it is possible to append
 units designator: \'K', \'M', \'G', \'T', \'P', or \'E', which represent
-KiB, MiB, GiB, TiB, PiB, or EiB, respectively. Case does not matter.
+KiB, MiB, GiB, TiB, PiB, or EiB, respectively (case does not matter).
++
+NOTE: Directory arguments without '-r' do not defragment files recursively but will
+defragment certain internal trees (extent tree and the subvolume tree). This has been
+confusing and could be removed in the future.
 
 *label* [<dev>|<mountpoint>] [<newlabel>]::
-Show or update the label of a filesystem.
+Show or update the label of a filesystem. This works on a mounted filesystem or
+a filesystem image.
 +
-[<device>|<mountpoint>] is used to identify the filesystem.
-If a newlabel optional argument is passed, the label is changed.
+The 'newlabel' argument is optional. Current label is printed if the the argument
+is omitted.
 +
-NOTE: the maximum allowable length shall be less than 256 chars
+NOTE: the maximum allowable length shall be less than 256 chars and must not contain
+a newline. The trailing newline is stripped automatically.
 
 // Some wording are extracted by the resize2fs man page
 *resize* [<devid>:][+/-]<size>[kKmMgGtTpPeE]|[<devid>:]max <path>::
-Resize a mounted filesystem identified by directory <path>. A particular device
-can be resized by specifying a <devid>.
+Resize a mounted filesystem identified by 'path'. A particular device
+can be resized by specifying a 'devid'.
 +
-If <path> is a file containing a btrfs image then resize does not work as
-expected and does not resize the image. This would resize the underlying
+WARNING: If 'path' is a file containing a BTRFS image then resize does not work
+as expected and does not resize the image. This would resize the underlying
 filesystem instead.
 +
-The devid can be found with *btrfs filesystem show* and
+The 'devid' can be found in the output of *btrfs filesystem show* and
 defaults to 1 if not specified.
-The <size> parameter specifies the new size of the filesystem.
-If the prefix + or - is present the size is increased or decreased
-by the quantity <size>.
-If no units are specified, the unit of the <size> parameter defaults to
-bytes. Optionally, the size parameter may be suffixed by one of the following
+The 'size' parameter specifies the new size of the filesystem.
+If the prefix '+' or '-' is present the size is increased or decreased
+by the quantity 'size'.
+If no units are specified, bytes are assumed for 'size'.
+Optionally, the size parameter may be suffixed by one of the following
 units designators: \'K', \'M', \'G', \'T', \'P', or \'E', which represent
-KiB, MiB, GiB, TiB, PiB, or EiB, respectively. Case does not matter.
+KiB, MiB, GiB, TiB, PiB, or EiB, respectively (case does not matter).
 +
-If \'max' is passed, the filesystem will occupy all available space on the
-device devid.
+If 'max' is passed, the filesystem will occupy all available space on the
+device respecting 'devid' (remember, devid 1 by default).
 +
 The resize command does not manipulate the size of underlying
 partition.  If you wish to enlarge/reduce a filesystem, you must make sure you
@@ -123,21 +164,29 @@ can expand the partition before enlarging the filesystem and shrink the
 partition after reducing the size of the filesystem.  This can done using
 `fdisk`(8) or `parted`(8) to delete the existing partition and recreate
 it with the new desired size.  When recreating the partition make sure to use
-the same starting disk cylinder as before.
+the same starting partition offset as before.
++
+Growing is usually instant as it only updates the size. However, shrinking could
+take a long time if there are data in the device area that's beyond the new
+end. Relocation of the data takes time.
++
+See also section 'EXAMPLES'.
 
 *show* [options] [<path>|<uuid>|<device>|<label>]::
-Show the btrfs filesystem with some additional info.
+Show the btrfs filesystem with some additional info about devices and space
+allocation.
 +
-If no option nor <path>|<uuid>|<device>|<label> is passed, btrfs shows
-information of all the btrfs filesystem both mounted and unmounted.
+If no option none of 'path'/'uuid'/'device'/'label' is passed, information
+about all the BTRFS filesystems is shown, both mounted and unmounted. 
 +
 `Options`
 +
 -m|--mounted::::
-probe btrfs kernel to list mounted btrfs filesystems(s)
+probe kernel for mounted BTRFS filesystems
 -d|--all-devices::::
 scan all devices under /dev, otherwise the devices list is extracted from the
-/proc/partitions file.
+/proc/partitions file. This is a fallback option if there's no device node
+manager (like udev) available in the system.
 --raw::::
 raw numbers in bytes, without the 'B' suffix
 --human-readable::::
@@ -156,10 +205,48 @@ show sizes in GiB, or GB with --si
 show sizes in TiB, or TB with --si
 
 *sync* <path>::
-Force a sync for the filesystem identified by <path>.
+Force a sync of the filesystem at 'path'. This is done via a special ioctl and
+will also trigger cleaning of deleted subvolumes. Besides that it's equivalent
+to the `sync`(1) command.
 
 *usage* [options] <path> [<path>...]::
-Show detailed information about internal filesystem usage.
+Show detailed information about internal filesystem usage. This is supposed to
+replace the *btrfs filesystem df* command in the long run.
++
+The level of detail can differ if the command is run under a regular or the
+root user (due to use of restricted ioctl). For both there's a summary section
+with information about space usage:
++
+-------------------------
+$ btrfs fi usage /path
+WARNING: cannot read detailed chunk info, RAID5/6 numbers will be incorrect, run as root
+Overall:
+    Device size:                   1.82TiB
+    Device allocated:              1.17TiB
+    Device unallocated:          669.99GiB
+    Device missing:                  0.00B
+    Used:                          1.14TiB
+    Free (estimated):            692.57GiB      (min: 692.57GiB)
+    Data ratio:                       1.00
+    Metadata ratio:                   1.00
+    Global reserve:              512.00MiB      (used: 0.00B)
+-------------------------
++
+The root user will also see stats broken down by block group types:
++
+-------------------------
+Data,single: Size:1.15TiB, Used:1.13TiB
+   /dev/sdb        1.15TiB
+
+Metadata,single: Size:12.00GiB, Used:6.45GiB
+   /dev/sdb       12.00GiB
+
+System,single: Size:32.00MiB, Used:144.00KiB
+   /dev/sdb       32.00MiB
+
+Unallocated:
+   /dev/sdb      669.99GiB
+-------------------------
 +
 `Options`
 +
@@ -186,6 +273,55 @@ show data in tabular format
 +
 If conflicting options are passed, the last one takes precedence.
 
+EXAMPLES
+--------
+
+*$ btrfs filesystem defrag -v -r dir/*
+
+Recursively defragment files under 'dir/', print files as they are processed.
+The file names will be printed in batches, similarly the amount of data triggered
+by defragmentation will be proportional to last N printed files. The system dirty
+memory throttling will slow down the defragmentation but there can still be a lot
+of IO load and the system may stall for a moment.
+
+*$ btrfs filesystem defrag -v -r -f dir/*
+
+Recusively defragment files under 'dir/', be verbose and wait until all blocks
+are flushed before processing next file. You can note slower progress of the
+output and lower IO load (proportional to currently defragmented file).
+
+*$ btrfs filesystem defrag -v -r -f -clzo dir/*
+
+Recusively defragment files under 'dir/', be verbose, wait until all blocks are
+flushed and force file compression.
+
+*$ btrfs filesystem defrag -v -r -t 64M dir/*
+
+Recusively defragment files under 'dir/', be verbose and try to merge extents
+to be about 64MiB. As stated above, the success rate depends on actual free
+space fragmentation and the final result is not guaranteed to meet the target
+even if run repeatedly.
+
+*$ btrfs filesystem resize -1G /path*
+
+*$ btrfs filesystem resize 1:-1G /path*
+
+Shrink size of the filesystem's device id 1 by 1GiB. The first syntax expects a
+device with id 1 to exist, otherwise fails. The second is equivalent and more
+explicit. For a single-device filesystem it's typically not necessary to
+specify the devid though.
+
+*$ btrfs filesystem resize max /path*
+
+*$ btrfs filesystem resize 1:max /path*
+
+Let's assume that devid 1 exists, the filesystem does not occupy the whole block
+device, eg. it has been enlarged and we wan the grow the filesystem. Simply using
+'max' as size we will achieve that.
+
+NOTE: There are two ways to minimize the filesystem on a given device. The
+*btrfs inspect-internal min-dev-size* command, or iteratively shrink in steps.
+
 EXIT STATUS
 -----------
 *btrfs filesystem* returns a zero exit status if it succeeds. Non zero is