If Borg is running in client/server mode, the client uses SSH as a transport to talk to the remote agent, which is another Borg process (Borg is installed on the server, too) started automatically by the client. The Borg server is doing storage-related low-level repo operations (get, put, commit, check, compact), while the Borg client does the high-level stuff: deduplication, encryption, compression, dealing with archives, backups, restores, etc., which reduces the amount of data that goes over the network.
When Borg is writing to a repo on a locally mounted remote file system, e.g. SSHFS, the Borg client only can do file system operations and has no agent running on the remote side, so every operation needs to go over the network, which is slower.
Yes, this is possible from the technical standpoint, but it is not recommended from the security perspective. BorgBackup is built upon a defined Attack model that cannot provide its guarantees for multiple clients using the same repository. See Isn’t BorgBackup’s AES-CTR crypto broken? for a detailed explanation.
Also, in order for the deduplication used by Borg to work, it
needs to keep a local cache containing checksums of all file
chunks already stored in the repository. This cache is stored in
~/.cache/borg/
. If Borg detects that a repository has been
modified since the local cache was updated it will need to rebuild
the cache. This rebuild can be quite time consuming.
So, yes it’s possible. But it will be most efficient if a single repository is only modified from one place. Also keep in mind that Borg will keep an exclusive lock on the repository while creating or deleting archives, which may make simultaneous backups fail.
It is possible to swap your backup disks if each backup medium is assigned its own repository by creating a new one with borg init.
If you want to have redundant backup repositories (preferably at separate locations), the recommended way to do that is like this:
borg init repo1
borg init repo2
client machine ---borg create---> repo1
client machine ---borg create---> repo2
This will create distinct repositories (separate repo ID, separate keys) and nothing bad happening in repo1 will influence repo2.
Some people decide against above recommendation and create identical copies of a repo (using some copy / sync / clone tool).
While this might be better than having no redundancy at all, you have to be very careful about how you do that and what you may / must not do with the result (if you decide against our recommendation).
What you would get with this is:
client machine ---borg create---> repo
repo ---copy/sync---> copy-of-repo
There is no special borg command to do the copying, you could just use any reliable tool that creates an identical copy (cp, rsync, rclone might be options).
But think about whether that is really what you want. If something goes wrong in repo, you will have the same issue in copy-of-repo.
Make sure you do the copy/sync while no backup is running, see borg with-lock about how to do that.
Also, you must not run borg against multiple instances of the same repo (like repo and copy-of-repo) as that would create severe issues:
Data loss: they have the same repository ID, so the borg client will think they are identical and e.g. use the same local cache for them (which is an issue if they happen to be not the same). See #4272 for an example.
Encryption security issues if you would update repo and copy-of-repo independently, due to AES counter reuse.
See also: My repository is corrupt, how can I restore from an older copy of it?
About the warning:
Cache, or information obtained from the security directory is newer than repository - this is either an attack or unsafe (multiple repos with same ID)
“unsafe”: If not following the advice from the previous section, you can easily run into this by yourself by restoring an older copy of your repository.
“attack”: maybe an attacker has replaced your repo by an older copy, trying to trick you into AES counter reuse, trying to break your repo encryption.
Borg users have also reported that fs issues (like hw issues / I/O errors causing the fs to become read-only) can cause this warning, see #7853.
If you’ld decide to ignore this and accept unsafe operation for this repository, you could delete the manifest-timestamp and the local cache:
borg config repo id # shows the REPO_ID
rm ~/.config/borg/security/REPO_ID/manifest-timestamp
borg delete --cache-only REPO
This is an unsafe and unsupported way to use borg, you have been warned.
UNIX domain sockets (because it does not make sense - they are meaningless without the running process that created them and the process needs to recreate them in any case). So, don’t panic if your backup misses a UDS!
The precise on-disk (or rather: not-on-disk) representation of the holes in a sparse file. Archive creation has no special support for sparse files, holes are backed up as (deduplicated and compressed) runs of zero bytes. Archive extraction has optional support to extract all-zero chunks as holes in a sparse file.
Some filesystem specific attributes, like btrfs NOCOW, see Support for file metadata.
For hardlinked symlinks, the hardlinking can not be archived (and thus, the hardlinking will not be done at extraction time). The symlinks will be archived and extracted as non-hardlinked symlinks, see #2379.
A single archive can only reference a limited volume of file/dir metadata, usually corresponding to tens or hundreds of millions of files/dirs. When trying to go beyond that limit, you will get a fatal IntegrityError exception telling that the (archive) object is too big. An easy workaround is to create multiple archives with fewer items each. See also the Note about archive limitations and #1452.
borg info shows how large (relative to the maximum size) existing archives are.
borg extract only supports restoring into an empty destination. After that, the destination will exactly have the contents of the extracted archive. If you extract into a non-empty destination, borg will (for example) not remove files which are in the destination, but not in the archive. See #4598 for a workaround and more details.
Yes, Borg supports resuming backups.
During a backup a special checkpoint archive named <archive-name>.checkpoint
is saved every checkpoint interval (the default value for this is 30
minutes) containing all the data backed-up until that point.
This checkpoint archive is a valid archive, but it is only a partial backup (not all files that you wanted to backup are contained in it). Having it in the repo until a successful, full backup is completed is useful because it references all the transmitted chunks up to the checkpoint. This means that in case of an interruption, you only need to retransfer the data since the last checkpoint.
If a backup was interrupted, you normally do not need to do anything special,
just invoke borg create
as you always do. If the repository is still locked,
you may need to run borg break-lock
before the next backup. You may use the
same archive name as in previous attempt or a different one (e.g. if you always
include the current datetime), it does not matter.
Borg always does full single-pass backups, so it will start again from the beginning - but it will be much faster, because some of the data was already stored into the repo (and is still referenced by the checkpoint archive), so it does not need to get transmitted and stored again.
Once your backup has finished successfully, you can delete all
<archive-name>.checkpoint
archives. If you run borg prune
, it will
also care for deleting unneeded checkpoints.
Note: the checkpointing mechanism creates hidden, partial files in an archive,
so that checkpoints even work while a big file is being processed.
They are named <filename>.borg_part_<N>
and all operations usually ignore
these files, but you can make them considered by giving the option
--consider-part-files
. You usually only need that option if you are
really desperate (e.g. if you have no completed backup of that file and you’ld
rather get a partial file extracted than nothing). You do not want to give
that option under any normal circumstances.
Note that checkpoints inside files are created only since version 1.1, make sure you have an up-to-date version of borgbackup if you want to continue instead of retransferring a huge file. In some cases, there is only an outdated version shipped with your distribution (e.g. Debian). See Installation.
This is not a problem anymore.
For more details, see If a backup stops mid-way, does the already-backed-up data stay there?.
You could do that (via borg config REPO append_only 0/1), but using different
ssh keys and different entries in authorized_keys
is much easier and also
maybe has less potential of things going wrong somehow.
When backing up your data over the network, your machine should not go to sleep. On macOS you can use caffeinate to avoid that.
If you cannot manage to extract the whole big file in one go, you can extract all the part files and manually concatenate them together.
For more details, see If a backup stops mid-way, does the already-backed-up data stay there?.
You can instruct export-tar
to send a tar stream to the stdout, and
then use tar
to perform the comparison:
borg export-tar /path/to/repo::archive-name - | tar --compare -f - -C /path/to/compare/to
If your repositories are encrypted and have the same ID, the recommended method is to delete the corrupted repository, but keep its security info, and then copy the working repository to the same location:
borg delete --keep-security-info /path/to/repo
rsync -aH /path/to/repo-working/ /path/to/repo # Note the trailing slash.
A plain delete command would remove the security info in
~/.config/borg/security
, including the nonce value. In BorgBackup
Encryption is AES-CTR, where the nonce is a counter. When the
working repo was used later for creating new archives, Borg would re-use nonce
values due to starting from a lower counter value given by the older copy of the
repository. To prevent this, the keep-security-info
option is applied so
that the client-side nonce counter is kept.
No, it can’t. While that at first sounds like a good idea to defend against some defect HDD sectors or SSD flash blocks, dealing with this in a reliable way needs a lot of low-level storage layout information and control which we do not have (and also can’t get, even if we wanted).
So, if you need that, consider RAID or a filesystem that offers redundant storage or just make backups to different locations / different hardware.
See also #225.
Yes, if you want to detect accidental data damage (like bit rot), use the
check
operation. It will notice corruption using CRCs and hashes.
If you want to be able to detect malicious tampering also, use an encrypted
repo. It will then be able to check using CRCs and HMACs.
SMR (shingled magnetic recording) hard drives are very different from regular hard drives. Applications have to behave in certain ways or performance will be heavily degraded.
Borg 1.1 ships with default settings suitable for SMR drives, and has been successfully tested on Seagate Archive v2 drives using the ext4 file system.
Some Linux kernel versions between 3.19 and 4.5 had various bugs handling device-managed SMR drives, leading to IO errors, unresponsive drives and unreliable operation in general.
For more details, refer to #2252.
A single error does not necessarily indicate bad hardware or a Borg bug. All hardware exhibits a bit error rate (BER). Hard drives are typically specified as exhibiting fewer than one error every 12 to 120 TB (one bit error in 10e14 to 10e15 bits). The specification is often called unrecoverable read error rate (URE rate).
Apart from these very rare errors there are two main causes of errors:
Defective hardware: described below.
Bugs in software (Borg, operating system, libraries): Ensure software is up to date. Check whether the issue is caused by any fixed bugs described in Important notes.
Finding defective hardware
Note
Hardware diagnostics are operating system dependent and do not apply universally. The commands shown apply for popular Unix-like systems. Refer to your operating system’s manual.
Find the drive containing the repository and use findmnt, mount or lsblk to learn the device path (typically /dev/…) of the drive. Then, smartmontools can retrieve self-diagnostics of the drive in question:
# smartctl -a /dev/sdSomething
The Offline_Uncorrectable, Current_Pending_Sector and Reported_Uncorrect attributes indicate data corruption. A high UDMA_CRC_Error_Count usually indicates a bad cable.
I/O errors logged by the system (refer to the system journal or dmesg) can point to issues as well. I/O errors only affecting the file system easily go unnoticed, since they are not reported to applications (e.g. Borg), while these errors can still corrupt data.
Drives can corrupt some sectors in one event, while remaining reliable otherwise. Conversely, drives can fail completely with no advance warning. If in doubt, copy all data from the drive in question to another drive -- just in case it fails completely.
If any of these are suspicious, a self-test is recommended:
# smartctl -t long /dev/sdSomething
Running fsck
if not done already might yield further insights.
Intermittent issues, such as borg check
finding errors
inconsistently between runs, are frequently caused by bad memory.
Run memtest86+ (or an equivalent memory tester) to verify that the memory subsystem is operating correctly.
Processors rarely cause errors. If they do, they are usually overclocked or otherwise operated outside their specifications. We do not recommend to operate hardware outside its specifications for productive use.
Tools to verify correct processor operation include Prime95 (mprime), linpack, and the Intel Processor Diagnostic Tool (applies only to Intel processors).
Repairing a damaged repository
With any defective hardware found and replaced, the damage done to the repository needs to be ascertained and fixed.
borg check provides diagnostics and --repair
options for repositories with
issues. We recommend to first run without --repair
to assess the situation.
If the found issues and proposed repairs seem right, re-run “check” with --repair
enabled.
If you noticed, there are some issues (#170 (warning: hell) and #4884) about the probability of a chunk having the same hash as another chunk, making the file corrupted because it grabbed the wrong chunk. This is called the Birthday Problem.
There is a lot of probability in here so, I can give you my interpretation of such math but it’s honestly better that you read it yourself and grab your own resolution from that.
Assuming that all your chunks have a size of \(2^{21}\) bytes (approximately 2.1 MB) and we have a “perfect” hash algorithm, we can think that the probability of collision would be of \(p^2/2^{n+1}\) then, using SHA-256 (\(n=256\)) and for example we have 1000 million chunks (\(p=10^9\)) (1000 million chunks would be about 2100TB). The probability would be around to 0.0000000000000000000000000000000000000000000000000000000000043.
A mass-murderer space rock happens about once every 30 million years on average. This leads to a probability of such an event occurring in the next second to about \(10^{-15}\). That’s 45 orders of magnitude more probable than the SHA-256 collision. Briefly stated, if you find SHA-256 collisions scary then your priorities are wrong. This example was grabbed from this SO answer, it’s great honestly.
Still, the real question is if Borg tries to not make this happen?
Well… it used to not check anything but there was a feature added which saves the size of the chunks too, so the size of the chunks is compared to the size that you got with the hash and if the check says there is a mismatch it will raise an exception instead of corrupting the file. This doesn’t save us from everything but reduces the chances of corruption. There are other ways of trying to escape this but it would affect performance so much that it wouldn’t be worth it and it would contradict Borg’s design, so if you don’t want this to happen, simply don’t use Borg.
Borg needs to write the time elapsed into the archive metadata before finalizing
the archive, compacting the segments, and committing the repo & cache. This means
when Borg is run with e.g. the time
command, the duration shown in the archive
stats may be shorter than the full time the command runs for.
Say you want to prune /var/log
faster than the rest of
/
. How do we implement that? The answer is to backup to different
archive names and then implement different prune policies for
different --glob-archives matching patterns.
For example, you could have a script that does:
borg create --exclude var/log $REPOSITORY:main-$(date +%Y-%m-%d) /
borg create $REPOSITORY:logs-$(date +%Y-%m-%d) /var/log
Then you would have two different prune calls with different policies:
borg prune --verbose --list -d 30 --glob-archives 'main-*' "$REPOSITORY"
borg prune --verbose --list -d 7 --glob-archives 'logs-*' "$REPOSITORY"
This will keep 7 days of logs and 30 days of everything else.
A file is only removed from a BorgBackup repository if all archives that contain the file are deleted and the corresponding data chunks are removed from the repository There are two ways how to remove files from a repository.
1. Use borg delete to remove all archives that contain the files. This will of course delete everything in the archive, not only some files.
2. If you really want to remove only some specific files, you can run the
borg recreate command to rewrite all archives with a different
--exclude
pattern. See the examples in the manpage for more information.
Finally, run borg compact with the --threshold 0
option to delete the
data chunks from the repository.
The compression level and algorithm don’t affect deduplication. Chunk ID hashes are calculated before compression. New compression settings will only be applied to new chunks, not existing chunks. So it’s safe to change them.
If a nonce (counter) value is reused, AES-CTR mode crypto is broken.
To exploit the AES counter management issue, an attacker would need to have access to the borg repository.
By tampering with the repo, the attacker could bring the repo into a state so that it reports a lower “highest used counter value” than the one that actually was used. The client would usually notice that, because it rather trusts the clientside stored “highest used counter value” than trusting the server.
But there are situations, where this is simply not possible:
If clients A and B used the repo, the client A can only know its own highest CTR value, but not the one produced by B. That is only known to (B and) the server (the repo) and thus the client A needs to trust the server about the value produced by B in that situation. You can’t do much about this except not having multiple clients per repo.
Even if there is only one client, if client-side information is completely lost (e.g. due to disk defect), the client also needs to trust the value from server side. You can avoid this by not continuing to write to the repository after you have lost clientside borg information.
The Borg config directory has content that you should take care of:
security
subdirectoryEach directory here represents one Borg repository by its ID and contains the last known status. If a repository’s status is different from this information at the beginning of BorgBackup operation, Borg outputs warning messages and asks for confirmation, so make sure you do not lose or manipulate these files. However, apart from those warnings, a loss of these files can be recovered.
keys
subdirectoryAll your borg keyfile keys are stored in this directory. Please note that borg repokey keys are stored inside the repository. You MUST make sure to have an independent backup of these keyfiles, otherwise you cannot access your backups anymore if you lose them. You also MUST keep these files secret; everyone who gains access to your repository and has the corresponding keyfile (and the key passphrase) can extract it.
Make sure that only you have access to the Borg config directory.
The cache contains a lot of metadata information about the files in your repositories and it is not encrypted.
However, the assumption is that the cache is being stored on the very same system which also contains the original files which are being backed up. So someone with access to the cache files would also have access the original files anyway.
The Internals section contains more details about The cache. If you ever need to move the cache to a different location, this can be achieved by using the appropriate Environment Variables.
There are several ways to specify a passphrase without human intervention:
BORG_PASSPHRASE
The passphrase can be specified using the BORG_PASSPHRASE
environment variable.
This is often the simplest option, but can be insecure if the script that sets it
is world-readable.
Note
Be careful how you set the environment; using the env
command, a system()
call or using inline shell scripts
(e.g. BORG_PASSPHRASE=hunter2 borg ...
)
might expose the credentials in the process list directly
and they will be readable to all users on a system. Using
export
in a shell script file should be safe, however, as
the environment of a process is accessible only to that
user.
BORG_PASSCOMMAND
with a properly permissioned fileAnother option is to create a file with a password in it in your home directory and use permissions to keep anyone else from reading it. For example, first create a key:
(umask 0077; head -c 32 /dev/urandom | base64 -w 0 > ~/.borg-passphrase)
Then in an automated script one can put:
export BORG_PASSCOMMAND="cat $HOME/.borg-passphrase"
and Borg will automatically use that passphrase.
It is possible to encrypt your repository in keyfile
mode instead of the default
repokey
mode and use a blank passphrase for the key file (simply press Enter twice
when borg init
asks for the password). See Repository encryption
for more details.
BORG_PASSCOMMAND
with macOS KeychainmacOS has a native manager for secrets (such as passphrases) which is safer
than just using a file as it is encrypted at rest and unlocked manually
(fortunately, the login keyring automatically unlocks when you login). With
the built-in security
command, you can access it from the command line,
making it useful for BORG_PASSCOMMAND
.
First generate a passphrase and use security
to save it to your login
(default) keychain:
security add-generic-password -D secret -U -a $USER -s borg-passphrase -w $(head -c 32 /dev/urandom | base64 -w 0)
In your backup script retrieve it in the BORG_PASSCOMMAND
:
export BORG_PASSCOMMAND="security find-generic-password -a $USER -s borg-passphrase -w"
BORG_PASSCOMMAND
with GNOME KeyringGNOME also has a keyring daemon that can be used to store a Borg passphrase.
First ensure libsecret-tools
, gnome-keyring
and libpam-gnome-keyring
are installed. If libpam-gnome-keyring
wasn’t already installed, ensure it
runs on login:
sudo sh -c "echo session optional pam_gnome_keyring.so auto_start >> /etc/pam.d/login"
sudo sh -c "echo password optional pam_gnome_keyring.so >> /etc/pam.d/passwd"
# you may need to relogin afterwards to activate the login keyring
Then add a secret to the login keyring:
head -c 32 /dev/urandom | base64 -w 0 | secret-tool store borg-repository repo-name --label="Borg Passphrase"
If a dialog box pops up prompting you to pick a password for a new keychain, use your login password. If there is a checkbox for automatically unlocking on login, check it to allow backups without any user intervention whatsoever.
Once the secret is saved, retrieve it in a backup script using BORG_PASSCOMMAND
:
export BORG_PASSCOMMAND="secret-tool lookup borg-repository repo-name"
Note
For this to automatically unlock the keychain it must be run
in the dbus
session of an unlocked terminal; for example, running a backup
script as a cron
job might not work unless you also export DISPLAY=:0
so secret-tool
can pick up your open session. It gets even more complicated
when you are running the tool as a different user (e.g. running a backup as root
with the password stored in the user keyring).
BORG_PASSCOMMAND
with KWalletKDE also has a keychain feature in the form of KWallet. The command-line tool
kwalletcli
can be used to store and retrieve secrets. Ensure kwalletcli
is installed, generate a passphrase, and store it in your “wallet”:
head -c 32 /dev/urandom | base64 -w 0 | kwalletcli -Pe borg-passphrase -f Passwords
Once the secret is saved, retrieve it in a backup script using BORG_PASSCOMMAND
:
export BORG_PASSCOMMAND="kwalletcli -e borg-passphrase -f Passwords"
Yes, file and directory metadata and data is locally encrypted, before leaving the local machine. We do not mean the transport layer encryption by that, but the data/metadata itself. Transport layer encryption (e.g. when ssh is used as a transport) applies additionally.
Yes and No.
No, as far as data confidentiality is concerned - if you use encryption, all your files/dirs data and metadata are stored in their encrypted form into the repository.
Yes, as an attacker with access to the remote server could delete (or otherwise make unavailable) all your backups.
Assume you backup your backup client machine C to the backup server S and C gets hacked. In a simple push setup, the attacker could then use borg on C to delete all backups residing on S.
These are your options to protect against that:
Do not allow to permanently delete data from the repo, see Append-only mode (forbid compaction).
Use a pull-mode setup using ssh -R
, see Backing up in pull mode for more information.
Mount C’s filesystem on another machine and then create a backup of it.
Do not give C filesystem-level access to S.
See Hosting repositories for a detailed protection guide.
Just in case you got the impression that pull-mode backups are way more safe than push-mode, you also need to consider the case that your backup server S gets hacked. In case S has access to a lot of clients C, that might bring you into even bigger trouble than a hacked backup client in the previous FAQ entry.
These are your options to protect against that:
Use the standard push-mode setup (see also previous FAQ entry).
Mount (the repo part of) S’s filesystem on C.
Do not give S file-system level access to C.
Have your backup server at a well protected place (maybe not reachable from the internet), configure it safely, apply security updates, monitor it, …
In general: if your only backup medium is nearby the backupped machine and always connected, you can easily get into trouble: they likely share the same fate if something goes really wrong.
Thus:
have multiple backup media
have media disconnected from network, power, computer
have media at another place
have a relatively recent backup on your media
Send a private email to the security contact if you think you have discovered a security issue. Please disclose security issues responsibly.
Borg uses AES-CTR encryption. An essential part of AES-CTR is a sequential counter that must never repeat. If the same value of the counter is used twice in the same repository, an attacker can decrypt the data. The counter is stored in the home directory of each user ($HOME/.config/borg/security/$REPO_ID/nonce) as well as in the repository (/path/to/repo/nonce). When creating a new archive borg uses the highest of the two values. The value of the counter in the repository may be higher than your local value if another user has created an archive more recently than you did.
Since the nonce is not necessary to read the data that is already encrypted,
borg info
, borg list
, borg extract
and borg mount
should work
just fine without it.
If the nonce file stored in the repo is lost, but you still have your local copy,
borg will recreate the repository nonce file the next time you run borg create
.
This should be safe for repositories that are only used from one user account
on one machine.
For repositories that are used by multiple users and/or from multiple machines it is safest to avoid running any commands that modify the repository after the nonce is deleted or if you suspect it may have been tampered with. See Attack model.
There can be many causes of this error. E.g. you have incorrectly specified the repository path.
You will also get this error if you try to access a repository that uses the argon2 key algorithm using an old version of borg. We recommend upgrading to the latest stable version and trying again. We are sorry. We should have thought abount forward compatibility and implemented a more helpful error message.
When I do a borg extract
, after a while all activity stops, no cpu usage,
no downloads.
This may happen when the SSH connection is stuck on server side. You can configure SSH on client side to prevent this by sending keep-alive requests, for example in ~/.ssh/config:
Host borg.example.com
# Client kills connection after 3*30 seconds without server response:
ServerAliveInterval 30
ServerAliveCountMax 3
You can also do the opposite and configure SSH on server side in /etc/ssh/sshd_config, to make the server send keep-alive requests to the client:
# Server kills connection after 3*30 seconds without client response:
ClientAliveInterval 30
ClientAliveCountMax 3
If you have issues with lost connections during long-running borg commands, you could try to work around:
Make partial extracts like borg extract REPO PATTERN
to do multiple
smaller extraction runs that complete before your connection has issues.
Try using borg mount REPO MOUNTPOINT
and rsync -avH
from
MOUNTPOINT
to your desired extraction directory. If the connection breaks
down, just repeat that over and over again until rsync does not find anything
to do any more. Due to the way borg mount works, this might be less efficient
than borg extract for bigger volumes of data.
When doing a backup to a remote server (using a ssh: repo URL), it sometimes stops after a while (some minutes, hours, … - not immediately) with “connection closed by remote” error message. Why?
That’s a good question and we are trying to find a good answer in #636.
Maybe the ssh connection between client and server broke down and that was not yet noticed on the server. Try these settings:
# /etc/ssh/sshd_config on borg repo server - kill connection to client
# after ClientAliveCountMax * ClientAliveInterval seconds with no response
ClientAliveInterval 20
ClientAliveCountMax 3
If you have multiple borg create … ; borg create … commands in a already
serialized way in a single script, you need to give them --lock-wait N
(with N
being a bit more than the time the server needs to terminate broken down
connections and release the lock).
This may especially happen if borg needs to rebuild the local “chunks” index - either because it was removed, or because it was not coherent with the repository state any more (e.g. because another borg instance changed the repository).
To optimize this rebuild process, borg caches per-archive information in the
chunks.archive.d/
directory. It won’t help the first time it happens, but it
will make the subsequent rebuilds faster (because it needs to transfer less data
from the repository). While being faster, the cache needs quite some disk space,
which might be unwanted.
There is a temporary (but maybe long lived) hack to avoid using lots of disk space for chunks.archive.d (see #235 for details):
# this assumes you are working with the same user as the backup.
cd ~/.cache/borg/$(borg config /path/to/repo id)
rm -rf chunks.archive.d ; touch chunks.archive.d
This deletes all the cached archive chunk indexes and replaces the directory that kept them with a file, so borg won’t be able to store anything “in” there in future.
This has some pros and cons, though:
much less disk space needs for ~/.cache/borg.
chunk cache resyncs will be slower as it will have to transfer chunk usage metadata for all archives from the repository (which might be slow if your repo connection is slow) and it will also have to build the hashtables from that data. chunk cache resyncs happen e.g. if your repo was written to by another machine (if you share same backup repo between multiple machines) or if your local chunks cache was lost somehow.
The long term plan to improve this is called “borgception”, see #474.
Backing up your entire root partition works just fine, but remember to
exclude directories that make no sense to backup, such as /dev, /proc,
/sys, /tmp and /run, and to use --one-file-system
if you only want to
backup the root partition (and not any mounted devices e.g.).
Check if your encoding is set correctly. For most POSIX-like systems, try:
export LANG=en_US.UTF-8 # or similar, important is correct charset
If that does not help:
check for typos, check if you really used export
.
check if you have set LC_ALL
- if so, try not setting it.
check if you generated the respective locale via locale-gen
.
This might be due to different ways to represent some characters in unicode or due to other non-ascii encoding issues.
If you run into that, try this:
avoid the non-ascii characters on the commandline by e.g. extracting the parent directory (or even everything)
mount the repo using FUSE and use some file manager
A first backup will usually be somehow “slow” because there is a lot of data to process. Performance here depends on a lot of factors, so it is hard to give specific numbers.
Subsequent backups are usually very fast if most files are unchanged and only a few are new or modified. The high performance on unchanged files primarily depends only on a few factors (like fs recursion + metadata reading performance and the files cache working as expected) and much less on other factors.
E.g., for this setup:
server grade machine (4C/8T 2013 Xeon, 64GB RAM, 2x good 7200RPM disks)
local zfs filesystem (mirrored) containing the backup source data
repository is remote (does not matter much for unchanged files)
backup job runs while machine is otherwise idle
The observed performance is that Borg can process about 1 million unchanged files (and a few small changed ones) in 4 minutes!
If you are seeing much less than that in similar circumstances, read the next few FAQ entries below.
So, if you feel your Borg backup is too slow somehow, you should find out why.
The usual way to approach this is to add --list --filter=AME --stats
to your
borg create
call to produce more log output, including a file list (with file status
characters) and also some statistics at the end of the backup.
Then you do the backup and look at the log output:
stats: Do you really have little changes or are there more changes than you thought? In the stats you can see the overall volume of changed data, which needed to be added to the repo. If that is a lot, that can be the reason why it is slow.
A
status (“added”) in the file list:
If you see that often, you have a lot of new files (files that Borg did not find
in the files cache). If you think there is something wrong with that (the file was there
already in the previous backup), please read the FAQ entries below.
M
status (“modified”) in the file list:
If you see that often, Borg thinks that a lot of your files might be modified
(Borg found them in the files cache, but the metadata read from the filesystem did
not match the metadata stored in the files cache).
In such a case, Borg will need to process the files’ contents completely, which is
much slower than processing unmodified files (Borg does not read their contents!).
The metadata values used in this comparison are determined by the --files-cache
option
and could be e.g. size, ctime and inode number (see the borg create
docs for more
details and potential issues).
You can use the stat
command on files to manually look at fs metadata to debug if
there is any unexpected change triggering the M
status.
Also, the --debug-topic=files_cache
option of borg create
provides a lot of debug
output helping to analyse why the files cache does not give its expected high performance.
When borg runs inside a virtual machine, there are some more things to look at:
Some hypervisors (e.g. kvm on proxmox) give some broadly compatible CPU type to the VM (usually to ease migration between VM hosts of potentially different hardware CPUs).
It is broadly compatible because they leave away modern CPU features that could be not present in older or other CPUs, e.g. hardware acceleration for AES crypto, for sha2 hashes, for (P)CLMUL(QDQ) computations useful for crc32.
So, basically you pay for compatibility with bad performance. If you prefer better performance, you should try to expose the host CPU’s misc. hw acceleration features to the VM which runs borg.
On Linux, check /proc/cpuinfo
for the CPU flags inside the VM.
For kvm check the docs about “Host model” and “Host passthrough”.
See also the next few FAQ entries for more details.
The files cache is used to determine whether Borg already “knows” / has backed up a file and if so, to skip the file from chunking. It intentionally excludes files that have a timestamp which is the same as the newest timestamp in the created archive.
So, if you see an ‘A’ status for unchanged file(s), they are likely the files with the most recent timestamp in that archive.
This is expected: it is to avoid data loss with files that are backed up from a snapshot and that are immediately changed after the snapshot (but within timestamp granularity time, so the timestamp would not change). Without the code that removes these files from the files cache, the change that happened right after the snapshot would not be contained in the next backup as Borg would think the file is unchanged.
This does not affect deduplication, the file will be chunked, but as the chunks will often be the same and already stored in the repo (except in the above mentioned rare condition), it will just re-use them as usual and not store new data chunks.
If you want to avoid unnecessary chunking, just create or touch a small or empty file in your backup source file set (so that one has the latest timestamp, not your 50GB VM disk image) and, if you do snapshots, do the snapshot after that.
Since only the files cache is used in the display of files status, those files are reported as being added when, really, chunks are already used.
By default, ctime (change time) is used for the timestamps to have a rather safe change detection (see also the --files-cache option).
Furthermore, pathnames recorded in files cache are always absolute, even if you specify source directories with relative pathname. If relative pathnames are stable, but absolute are not (for example if you mount a filesystem without stable mount points for each backup or if you are running the backup from a filesystem snapshot whose name is not stable), borg will assume that files are different and will report them as ‘added’, even though no new chunks will be actually recorded for them. To avoid this, you could bind mount your source directory in a directory with the stable path.
Borg maintains a files cache where it remembers the timestamp, size and inode of files. When Borg does a new backup and starts processing a file, it first looks whether the file has changed (compared to the values stored in the files cache). If the values are the same, the file is assumed unchanged and thus its contents won’t get chunked (again).
Borg can’t keep an infinite history of files of course, thus entries in the files cache have a “maximum time to live” which is set via the environment variable BORG_FILES_CACHE_TTL (and defaults to 20). Every time you do a backup (on the same machine, using the same user), the cache entries’ ttl values of files that were not “seen” are incremented by 1 and if they reach BORG_FILES_CACHE_TTL, the entry is removed from the cache.
So, for example, if you do daily backups of 26 different data sets A, B, C, …, Z on one machine (using the default TTL), the files from A will be already forgotten when you repeat the same backups on the next day and it will be slow because it would chunk all the files each time. If you set BORG_FILES_CACHE_TTL to at least 26 (or maybe even a small multiple of that), it would be much faster.
Besides using a higher BORG_FILES_CACHE_TTL (which also increases memory usage), there is also BORG_FILES_CACHE_SUFFIX which can be used to have separate (smaller) files caches for each backup set instead of the default one (big) unified files cache.
Another possible reason is that files don’t always have the same path, for example if you mount a filesystem without stable mount points for each backup or if you are running the backup from a filesystem snapshot whose name is not stable. If the directory where you mount a filesystem is different every time, Borg assumes they are different files. This is true even if you backup these files with relative pathnames - borg uses full pathnames in files cache regardless.
It is possible for some filesystems, such as mergerfs
or network filesystems,
to return inconsistent inode numbers across runs, causing borg to consider them changed.
A workaround is to set the option --files-cache=ctime,size
to exclude the inode
number comparison from the files cache check so that files with different inode
numbers won’t be treated as modified.
To limit upload (i.e. borg create) bandwidth, use the
--remote-ratelimit
option.
There is no built-in way to limit download (i.e. borg extract) bandwidth, but limiting download bandwidth can be accomplished with pipeviewer:
Create a wrapper script: /usr/local/bin/pv-wrapper
#!/bin/sh
## -q, --quiet do not output any transfer information at all
## -L, --rate-limit RATE limit transfer to RATE bytes per second
RATE=307200
pv -q -L $RATE | "$@"
Add BORG_RSH environment variable to use pipeviewer wrapper script with ssh.
export BORG_RSH='/usr/local/bin/pv-wrapper ssh'
Now Borg will be bandwidth limited. The nice thing about pv
is that you can
change rate-limit on the fly:
pv -R $(pidof pv) -L 102400
Possible use cases:
Another file system is mounted and you want to backup it with original paths.
You have created a BTRFS snapshot in a /.snapshots
directory for backup.
To achieve this, run borg create
within the mountpoint/snapshot directory:
# Example: Some file system mounted in /mnt/rootfs.
cd /mnt/rootfs
borg create /path/to/repo::rootfs_backup .
Borg is doing nothing special in the filesystem, it only uses very common and compatible operations (even the locking is just “rename”).
So, if you are encountering issues like slowness, corruption or malfunction when using a specific filesystem, please try if you can reproduce the issues with a local (non-network) and proven filesystem (like ext4 on Linux).
If you can’t reproduce the issue then, you maybe have found an issue within the filesystem code you used (not with Borg). For this case, it is recommended that you talk to the developers / support of the network fs and maybe open an issue in their issue tracker. Do not file an issue in the Borg issue tracker.
If you can reproduce the issue with the proven filesystem, please file an issue in the Borg issue tracker about that.
Repair usually works for recovering data in a corrupted archive. However, it’s impossible to predict all modes of corruption. In some very rare instances, such as malfunctioning storage hardware, additional repo corruption may occur. If you can’t afford to lose the repo, it’s strongly recommended that you perform repair on a copy of the repo.
Will perform automated routines that modify your backup repository
Might not actually fix the problem you are experiencing
Might, in very rare cases, further corrupt your repository
In the case of malfunctioning hardware, such as a drive or USB hub corrupting data when read or written, it’s best to diagnose and fix the cause of the initial corruption before attempting to repair the repo. If the corruption is caused by a one time event such as a power outage, running borg check --repair will fix most problems.
Some borg runs take quite a bit, so it would be nice to see a progress display, maybe even including a ETA (expected time of “arrival” [here rather “completion”]).
For some functionality, this can be done: if the total amount of work is more or
less known, we can display progress. So check if there is a --progress
option.
But sometimes, the total amount is unknown (e.g. for borg create
we just do
a single pass over the filesystem, so we do not know the total file count or data
volume before reaching the end). Adding another pass just to determine that would
take additional time and could be incorrect, if the filesystem is changing.
Even if the fs does not change and we knew count and size of all files, we still
could not compute the borg create
ETA as we do not know the amount of changed
chunks, how the bandwidth of source and destination or system performance might
fluctuate.
You see, trying to display ETA would be futile. The borg developers prefer to rather not implement progress / ETA display than doing futile attempts.
See also: https://xkcd.com/612/
By default, sshfs
is not entirely POSIX-compliant when renaming files due to
a technicality in the SFTP protocol. Fortunately, it also provides a workaround
to make it behave correctly:
sshfs -o workaround=rename user@host:dir /mnt/dir
In some cases, the free disk space of the target volume is reported incorrectly. This can happen for CIFS- or FUSE shares. If you are sure that your target volume will always have enough disk space, you can use the following workaround to disable checking for free disk space:
borg config -- $REPO_LOCATION additional_free_space -2T
There is nothing special that needs to be done, you can simply rename the
directory that corresponds to the repository. However, the next time borg
interacts with the repository (i.e, via borg list
), depending on the value
of BORG_RELOCATED_REPO_ACCESS_IS_OK
, borg may warn you that the repository
has been moved. You will be given a prompt to confirm you are OK with this.
If BORG_RELOCATED_REPO_ACCESS_IS_OK
is unset, borg will interactively ask for
each repository whether it’s OK.
It may be useful to set BORG_RELOCATED_REPO_ACCESS_IS_OK=yes
to avoid the
prompts when renaming multiple repositories or in a non-interactive context
such as a script. See Deployment for an example.
The simplest solution is to increase or disable the quota and resume the backup:
borg config /path/to/repo storage_quota 0
If you are bound to the quota, you have to free repository space. The first to try is running borg compact to free unused backup space (see also Separate compaction):
borg compact /path/to/repo
If your repository is already compacted, run borg prune or
borg delete to delete archives that you do not need anymore, and then run
borg compact
again.
Borg cannot work if you really have zero free space on the backup disk, so the first thing you must do is deleting some files to regain free disk space. See Important note about free space for further details.
Some Borg commands that do not change the repository might work under disk-full conditions, but generally this should be avoided. If your backup disk is already full when Borg starts a write command like borg create, it will abort immediately and the repository will stay as-is.
If you run a backup that stops due to a disk running full, Borg will roll back, delete the new new segment file and thus freeing disk space automatically. There may be a checkpoint archive left that has been saved before the disk got full. You can keep it to speed up the next backup or delete it to get back more disk space.
https://github.com/osxfuse/osxfuse/wiki/Mount-options#local
Read the above first and use this on your own risk:
borg mount -olocal REPO MOUNTPOINT
We try to build the binary on old, but still supported systems - to keep the minimum requirement for the (g)libc low. The (g)libc can’t be bundled into the binary as it needs to fit your kernel and OS, but Python and all other required libraries will be bundled into the binary.
If your system fulfills the minimum (g)libc requirement (see the README that is released with the binary), there should be no problem. If you are slightly below the required version, maybe just try. Due to the dynamic loading (or not loading) of some shared libraries, it might still work depending on what libraries are actually loaded and used.
In the borg git repository, there is scripts/glibc_check.py that can determine (based on the symbols’ versions they want to link to) whether a set of given (Linux) binaries works with a given glibc version.
Borg was created in May 2015 in response to the difficulty of getting new code or larger changes incorporated into Attic and establishing a bigger developer community / more open development.
More details can be found in ticket 217 that led to the fork.
Borg intends to be:
simple:
as simple as possible, but no simpler
do the right thing by default, but offer options
open:
welcome feature requests
accept pull requests of good quality and coding style
give feedback on PRs that can’t be accepted “as is”
discuss openly, don’t work in the dark
changing:
Borg is not compatible with Attic
do not break compatibility accidentally, without a good reason or without warning. allow compatibility breaking for other cases.
if major version number changes, it may have incompatible changes
Borg is a fork of Attic and maintained by “The Borg collective”.
Here’s a (incomplete) list of some major changes:
lots of attic issues fixed (see issue #5), including critical data corruption bugs and security issues.
more open, faster paced development (see issue #1)
less chunk management overhead (less memory and disk usage for chunks index)
faster remote cache resync (useful when backing up multiple machines into same repo)
compression: no, lz4, zstd, zlib or lzma compression, adjustable compression levels
repokey replaces problematic passphrase mode (you can’t change the passphrase nor the pbkdf2 iteration count in “passphrase” mode)
simple sparse file support, great for virtual machine disk files
can read special files (e.g. block devices) or from stdin, write to stdout
rename-based locking is more compatible than attic’s posix locking
uses fadvise to not spoil / blow up the fs cache
better error messages / exception handling
better logging, screen output, progress indication
tested on misc. Linux systems, 32 and 64bit, FreeBSD, OpenBSD, NetBSD, macOS
Please read the Change Log (or docs/changes.rst
in the source distribution) for more
information.
Borg is not compatible with original Attic (but there is a one-way conversion).
Use borg upgrade. This is a one-way process that cannot be reversed.
There are some caveats:
The upgrade can only be performed on local repositories. It cannot be performed on remote repositories.
If the repository is in “keyfile” encryption mode, the keyfile must exist locally or it must be manually moved after performing the upgrade:
Get the repository ID with borg config /path/to/repo id
.
Locate the attic key file at ~/.attic/keys/
. The correct key for the
repository starts with the line ATTIC_KEY <repository id>
.
Copy the attic key file to ~/.config/borg/keys/
Change the first line from ATTIC_KEY ...
to BORG_KEY ...
.
Verify that the repository is now accessible (e.g. borg list <repository>
).
Attic and Borg use different “chunker params”. This means that data added by Borg won’t deduplicate with the existing data stored by Attic. The effect is lessened if the files cache is used with Borg.
Repositories in “passphrase” mode must be migrated to “repokey” mode using Upgrading a passphrase encrypted attic repo. Borg does not support the “passphrase” mode any other way.
Attic was rather unflexible when it comes to compression, it always compressed using zlib level 6 (no way to switch compression off or adjust the level or algorithm).
The default in Borg is lz4, which is fast enough to not use significant CPU time in most cases, but can only achieve modest compression. It still compresses easily compressed data fairly well.
Borg also offers zstd, zlib and lzma compression, choose wisely.
Which choice is the best option depends on a number of factors, like bandwidth to the repository, how well the data compresses, available CPU power and so on.