BTRFS-BALANCE(8)

📦
version 7.3 · 20 versions
, in arch rolling

✨ BTRFS-BALANCE(8) ✨

okay bestie, gather round, because this is the btrfs-balance page from the btrfs manual (section 8, system administration) and it is a whole journey. grab a snack. 🍓

💫 NAME 💫

btrfs-balance — balance block groups on a btrfs filesystem

no cap, that's the whole vibe. it's the glow-up your block groups get when they're not where they're supposed to be. 💅

🌸 SYNOPSIS 🌸

btrfs balance <subcommand> <args>

🦋 DESCRIPTION 🦋

so the main character reason balance exists at all: it spreads your block groups across all the devices so they actually match the constraints set by their respective profiles. want the deep lore on profiles? go read mkfs.btrfs(8), section PROFILES. 📖

you can also tune how far-reaching the balance is by using filters, which pick and choose which block groups get processed. balance only works on a filesystem that's mounted — it's not doing anything while the fs is sitting there unmounted, bestie.

here's the good news ✨:

  • extent sharing is preserved — reflinks do not get broken
  • files are not defragmented and not recompressed
  • file extents are preserved

but the physical location on the devices? yeah, that will change. it's giving moving day, not a makeover.

balance is cancellable by you, the user. and the on-disk state of the filesystem is always consistent, so even if something goes horribly wrong — system crash, unexpected reboot, whatever — the filesystem does not get corrupted. the progress of the balance is temporarily stashed as an internal state and will be resumed upon mount, unless you passed the mount option skip_balance. 🫶

Warning: running balance without filters will take a lot of time. like, a lot a lot. it basically moves data/metadata from the whole entire filesystem and needs to update all block pointers. this is the "i have nothing else to do today" option. 😭

the filters can be used to do these things:

  • convert block group profiles (filter convert)
  • make block group usage more compact (filter usage)
  • perform actions only on a given device (filters devid, drange)

filters can be applied to a combination of block group types (data, metadata, system). but heads up ⚠️ — if you're changing only the system type, you need the force option. otherwise system gets automatically converted whenever the metadata profile is converted. it just tags along, iykyk.

and when metadata redundancy is being reduced (say, going from RAID1 to single), the force option is also required — and this gets noted in the system log. 📝

Note: the balance operation needs enough work space, i.e. space that is completely unused in the filesystem. if there isn't enough, you may get ENOSPC reports. see the section ENOSPC for more details. 👑

🌈 COMPATIBILITY 🌈

Note: the balance subcommand also exists under the btrfs filesystem namespace. it still works for backward compatibility but it is deprecated and should not be used any more. we don't do that here. 🙈
Note: a short syntax btrfs balance <path> works due to backward compatibility but is deprecated and should not be used any more. use the btrfs balance start command instead. periodt. 💅

🔥 PERFORMANCE IMPLICATIONS 🔥

balancing operations are very IO intensive and can also be pretty CPU intensive, which means they will impact other filesystem operations happening at the same time. typically, huge amounts of data get copied from one location to another, with the corresponding metadata updates trailing behind like a group chat. 💀

depending on the block group layout, it can also be seek heavy. and performance on rotational devices is noticeably worse compared to SSDs or fast arrays. spinning rust really said "not today". 🐌

👑 SUBCOMMAND 👑

cancel <path>

cancels a running or paused balance. the command will block and wait until the current block group being processed completes.

since kernel 5.7 the response time of the cancellation is significantly improved. on older kernels it might take a long time until the currently processed chunk is completely finished. patience, bestie. 🧘

pause <path>

pause a running balance operation. this will store the state of the balance progress and the used filters to the filesystem. 💾

resume <path>

resume an interrupted balance. the balance status must be stored on the filesystem from a previous run — e.g. after it was paused, or forcibly interrupted and then mounted again with skip_balance. 🫶

start [options] <path>

start the balance operation according to the specified filters. without any filters, the data and metadata from the whole filesystem are moved. the process runs in the foreground.

Note: the balance command without filters will basically move everything in the filesystem to a new physical location on devices (i.e. it does not affect the logical properties of file extents like offsets within files and extent sharing). the run time is potentially very long, depending on the filesystem size. to prevent starting a full balance by accident, the user is warned and has a few seconds to cancel the operation before it starts. the warning and the delay can be skipped with the --full-balance option. ⏳

please note that the filters must be written together with the -d, -m and -s options, because they're optional and bare -d and -m also work and mean no filters. don't leave your filters hanging, they need their letters. 🎀

Note: when the target profile for the conversion filter is raid5 or raid6, there's a safety timeout of 10 seconds to warn users about the status of the feature. ⏰

Options

  • -d[<filters>] — act on data block groups, see section FILTERS for details about filters
  • -m[<filters>] — act on metadata chunks, see FILTERS for details about filters
  • -s[<filters>] — act on system chunks (requires -f), see FILTERS for details about filters.
  • -f — force a reduction of metadata integrity, e.g. when going from raid1 to single, or skip the safety timeout when the target conversion profile is raid5 or raid6
  • --background|--bg — run the balance operation asynchronously in the background, uses fork(2) to start the process that calls the kernel ioctl
  • --enqueue — wait if there's another exclusive operation running, otherwise continue
  • -v — (deprecated) alias for the global -v option

status [-v] <path>

Show status of running or paused balance. 👀

Options

  • -v — (deprecated) alias for the global -v option

✨ FILTERS ✨

okay so from kernel 3.3 onwards, BTRFS balance can limit its action to just a subset of the whole filesystem, and you can use it to change up your replication configuration (like converting data from single to RAID1). 💅

Balance can be limited to a block group profile with these options:

  • -d — for data block groups
  • -m — for metadata block groups (also implicitly applies to -s)
  • -s — for system block groups

These options have an optional parameter, which means the parameter must start right after the option without a space (it's mandatory getopt syntax, bestie), like -dusage=10. You can specify options for all block group types in one command. slay.

A filter has this structure: filter[=params][,filter=...]

To combine multiple filters use ,, without spaces. Example: -dconvert=raid1,soft. Note that most filters can only be specified once per block group profile (data/metadata/system block groups).

BTRFS can have different profiles on a single device or the same profile on multiple device.

The main reason you'd want different profiles for data and metadata is to give extra protection to the filesystem's metadata when devices fail. Like, a single sector of unrecoverable metadata will break the filesystem, while a single sector of lost data can be trivially recovered by deleting the broken file. 💀

Before changing profiles, make sure there's enough unallocated space on existing drives to create new metadata block groups (for filesystems over 50GiB, this is 1GB * (number_of_devices + 2)).

Default profiles on BTRFS are:

  • data: single
  • metadata:
    • single devices: dup
    • multiple devices: raid1

The available filter types are:

✨ Filter types ✨

  • profiles=<profiles> — Balances only block groups with the given profiles. Parameters are a list of profile names separated by | (pipe). 🎀
  • usage=<percent>, usage=<range> — Balances only block groups with usage under the given percentage. The value 0 is allowed and will clean up completely unused block groups, and this shouldn't require any new work space allocated. You might wanna use usage=0 if balance is returning ENOSPC and your filesystem isn't too full.

The argument can be a single value or a range. A single value N means at most N percent used, equivalent to ..N range syntax. Kernels before 4.4 only accept the single value format. The minimum range boundary is inclusive, maximum is exclusive.

This filter can only be specified once per block group profile.

  • devid=<id> — Balances only block groups which have at least one chunk on the given device. To list devices with ids use btrfs filesystem show.

This filter can only be specified once per block group profile.

  • drange=<range> — Balances only block groups which overlap with the given byte range on any device. Use this in conjunction with devid to filter on a specific device. The parameter is a range specified as start..end.

This filter can only be specified once per block group profile.

  • vrange=<range> — Balances only block groups which overlap with the given byte range in the filesystem's internal virtual address space. This is the address space that most reports from btrfs in the kernel log use. The parameter is a range specified as start..end.

This filter can only be specified once per block group profile.

  • convert=<profile> — Convert each selected block group to the given profile name identified by parameters.

This filter can only be specified once per block group profile.

Note: Starting with kernel 4.5, data chunks can be converted to/from the DUP profile on a single device. Starting with kernel 4.6, all profiles can be converted to/from DUP on multi-device filesystems.

Warning: Bad or missing devices are not detected immediately during runtime, and this depends on some later event like a failed write or failed transaction commit. If there's a known failing device, or a device deleted by the /sys/block/<dev>/device/delete interface, the device will still be accessed and written to. In such a case, you should not convert to a profile with lower redundancy (e.g. from RAID1 to SINGLE), as attempts to create new chunks on the new devices will cause various problems. The proper action is to use btrfs replace or btrfs device remove to handle the failing/missing device first. Then convert will work with all devices correctly.
  • limit=<number>, limit=<range> — Process only the given number of chunks, after all filters are applied. This can be used to specifically target a chunk in connection with other filters (drange, vrange) or just simply limit the amount of work done by a single balance run.

The argument can be a single value or a range. A single value N means at most N chunks, equivalent to ..N range syntax. Kernels before 4.4 only accept the single value format. The range minimum and maximum are inclusive.

This filter can only be specified once per block group profile.

  • stripes=<range> — Balances only block groups which have the given number of stripes. The parameter is a range specified as start..end. Makes sense for block group profiles that utilize striping, i.e. RAID0/10/5/6. The range minimum and maximum are inclusive.

This filter can only be specified once per block group profile.

  • soft — Takes no parameters. Only has meaning when converting between profiles. When doing a convert from one profile to another and soft mode is on, chunks that already have the target profile are left untouched. This is useful e.g. when half of the filesystem was converted earlier but got cancelled.

The soft mode switch is (like every other filter) per-type. For example, this means we can convert metadata chunks the "hard" way while converting data chunks selectively with the soft switch.

This filter can only be specified once per block group profile.

Profile names, used in profiles and convert, are one of:

  • raid0
  • raid1
  • raid1c3
  • raid1c4
  • raid10
  • raid5
  • raid6
  • dup
  • single

The mixed data/metadata profiles can be converted in the same way, but conversion between mixed and non-mixed is not implemented. For the constraints of the profiles please refer to mkfs.btrfs(8) section PROFILES.

💖 ENOSPC 💖

The way balance operates, it usually needs to temporarily create a new block group and move the old data there, before the old block group can be removed. For that it needs the work space, otherwise it fails for ENOSPC reasons. This is not the same ENOSPC as if the free space is exhausted. This refers to the space on the level of block groups, which are bigger parts of the filesystem that contain many file extents.

The free work space can be calculated from the output of the btrfs filesystem show command:

   Label: 'BTRFS'  uuid: 8a9d72cd-ead3-469d-b371-9c7203276265
           Total devices 2 FS bytes used 77.03GiB
           devid    1 size 53.90GiB used 51.90GiB path /dev/sdc2
           devid    2 size 53.90GiB used 51.90GiB path /dev/sde1

size - used = free work space

53.90GiB - 51.90GiB = 2.00GiB

An example of a filter that does not require workspace is usage=0. This will scan through all unused block groups of a given type and will reclaim the space. After that it might be possible to run other filters.

🦋 CONVERSIONS ON MULTIPLE DEVICES 🦋

Conversion to profiles based on striping (RAID0, RAID5/6) require the work space on each device. An interrupted balance may leave partially filled block groups that consume the work space.

✨ EXAMPLES ✨

If you want the full glow-up story — going from one device to multiple devices and then back again — there's a whole more comprehensive example waiting for you over in the TYPICAL USECASES section of btrfs-device(8) <>. bestie, it's giving ✨complete saga✨ 🦋

🌸 MAKING BLOCK GROUP LAYOUT MORE COMPACT 🌸

Okay so real talk: the layout of block groups is not normally visible 🙈 most tools only report summarized numbers of free or used space, but there are still some hints provided. iykyk.

Let's use the following real life example and start with the output:

$ btrfs filesystem df /path
Data, single: total=75.81GiB, used=64.44GiB
System, RAID1: total=32.00MiB, used=20.00KiB
Metadata, RAID1: total=15.87GiB, used=8.84GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

Roughly calculating for data, 75G - 64G = 11G, the used/total ratio is about 85%. So how can we interpret that, you ask? 👑

  • chunks are filled by 85% on average, i.e. the usage filter with anything smaller than 85 will likely not affect anything
  • in a more realistic scenario, the space is distributed unevenly, we can assume there are completely used chunks and the remaining are partially filled

Compacting the layout could be used on both. In the former case it would spread data of a given chunk to the others and removing it. Here we can estimate that roughly 850 MiB of data have to be moved (85% of a 1 GiB chunk). 💅

In the latter case, targeting the partially used chunks will have to move less data and thus will be faster 🔥 A typical filter command would look like:

# btrfs balance start -dusage=50 /path
Done, had to relocate 2 out of 97 chunks

$ btrfs filesystem df /path
Data, single: total=74.03GiB, used=64.43GiB
System, RAID1: total=32.00MiB, used=20.00KiB
Metadata, RAID1: total=15.87GiB, used=8.84GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

As you can see, the total amount of data is decreased by just 1 GiB, which is an expected result. Now let's see what will happen when we increase the estimated usage filter 👀

# btrfs balance start -dusage=85 /path
Done, had to relocate 13 out of 95 chunks

$ btrfs filesystem df /path
Data, single: total=68.03GiB, used=64.43GiB
System, RAID1: total=32.00MiB, used=20.00KiB
Metadata, RAID1: total=15.87GiB, used=8.85GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

Now the used/total ratio is about 94% and we moved about 74G - 68G = 6G of data to the remaining block groups, i.e. the 6GiB are now free of filesystem structures, and can be reused for new data or metadata block groups. ate and left no crumbs 🍓

We can do a similar exercise with the metadata block groups, but this should not typically be necessary, unless the used/total ratio is really off. Here the ratio is roughly 50% but the difference as an absolute number is "a few gigabytes", which can be considered normal for a workload with snapshots or reflinks updated frequently.

# btrfs balance start -musage=50 /path
Done, had to relocate 4 out of 89 chunks

$ btrfs filesystem df /path
Data, single: total=68.03GiB, used=64.43GiB
System, RAID1: total=32.00MiB, used=20.00KiB
Metadata, RAID1: total=14.87GiB, used=8.85GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

Just 1 GiB decrease, which possibly means there are block groups with good utilization. Making the metadata layout more compact would in turn require updating more metadata structures, i.e. lots of IO 😭 As running out of metadata space is a more severe problem, it's not necessary to keep the utilization ratio too high. For the purpose of this example, let's see the effects of further compaction:

# btrfs balance start -musage=70 /path
Done, had to relocate 13 out of 88 chunks

$ btrfs filesystem df .
Data, single: total=68.03GiB, used=64.43GiB
System, RAID1: total=32.00MiB, used=20.00KiB
Metadata, RAID1: total=11.97GiB, used=8.83GiB
GlobalReserve, single: total=512.00MiB, used=0.00B

🎀 GETTING RID OF COMPLETELY UNUSED BLOCK GROUPS 🎀

Normally the balance operation needs a work space, to temporarily move the data before the old block groups gets removed. If there's no work space, it ends with no space left 💀

There's a special case when the block groups are completely unused, possibly left after removing lots of files or deleting snapshots. Removing empty block groups is automatic since 3.18 ⭐ The same can be achieved manually with a notable exception that this operation does not require the work space. Thus it can be used to reclaim unused block groups to make it available.

# btrfs balance start -dusage=0 /path

This should lead to decrease in the total numbers in the btrfs filesystem df output. no cap 📉

💖 EXIT STATUS 💖

Unless indicated otherwise below, all btrfs balance subcommands return a zero exit status if they succeed, and non zero in case of failure.

The pause, cancel, and resume subcommands exit with a status of 2 if they fail because a balance operation was not running.

The status subcommand exits with a status of 0 if a balance operation is not running, 1 if the command-line usage is incorrect or a balance operation is still running, and 2 on other errors. note it down bestie 🫶

🌈 AVAILABILITY 🌈

btrfs is part of btrfs-progs. Please refer to the documentation at <https://btrfs.readthedocs.io>. 💜

👑 SEE ALSO 👑

mkfs.btrfs(8) <>, btrfs-device(8) <>

and that's a wrap on the whole page, bestie 🎀 this was btrfs-progs version 7.3, dated Oct 24, 2026 — go compact some block groups, you absolute icon 💅✨🦋

📦 where this page comes from