Releases: restic/restic
restic 0.18.1
Changelog for restic 0.18.1 (2025-09-21)
The following sections list the changes in restic 0.18.1 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #5324: Correctly handle
backup --stdin-filenamewith directory paths - Fix #5325: Accept
RESTIC_HOSTenvironment variable inforgetcommand - Fix #5342: Ignore "chmod not supported" errors when writing files
- Fix #5344: Ignore
EOPNOTSUPPerrors for extended attributes - Fix #5421: Fix rare crash if directory is removed during backup
- Fix #5429: Stop retrying uploads when rest-server runs out of space
- Fix #5467: Improve handling of download retries in
checkcommand
Details
-
Bugfix #5324: Correctly handle
backup --stdin-filenamewith directory pathsIn restic 0.18.0, the
backupcommand failed if a filename that includes at least a directory was passed to--stdin-filename. For example,--stdin-filename /foo/barresulted in the following error:Fatal: unable to save snapshot: open /foo: no such file or directoryThis has now been fixed.
-
Bugfix #5325: Accept
RESTIC_HOSTenvironment variable inforgetcommandThe
forgetcommand did not use the host name from theRESTIC_HOSTenvironment variable when filtering snapshots. This has now been fixed. -
Bugfix #5342: Ignore "chmod not supported" errors when writing files
Restic 0.18.0 introduced a bug that caused
chmod xxx: operation not supportederrors to appear when writing to a local file repository that did not support chmod (like CIFS or WebDAV mounted via FUSE). Restic now ignores those errors. -
Bugfix #5344: Ignore
EOPNOTSUPPerrors for extended attributesRestic 0.18.0 added extended attribute support for NetBSD 10+, but not all NetBSD filesystems support extended attributes. Other BSD systems can likewise return
EOPNOTSUPP, so restic now ignores these errors. -
Bugfix #5421: Fix rare crash if directory is removed during backup
In restic 0.18.0, the
backupcommand could crash if a directory was removed between reading its metadata and listing its directory content. This has now been fixed. -
Bugfix #5429: Stop retrying uploads when rest-server runs out of space
When rest-server returns a
507 Insufficient Storageerror, it indicates that no more storage capacity is available. Restic now correctly stops retrying uploads in this case. -
Bugfix #5467: Improve handling of download retries in
checkcommandIn very rare cases, the
checkcommand could unnecessarily report repository damage if the backend returned incomplete, corrupted data on the first download try which is afterwards resolved by a download retry.This could result in an error output like the following:
Load(<data/34567890ab>, 33918928, 0) returned error, retrying after 871.35598ms: readFull: unexpected EOF Load(<data/34567890ab>, 33918928, 0) operation successful after 1 retries check successful on second attempt, original error pack 34567890ab[...] contains 6 errors: [blob 12345678[...]: decrypting blob <data/12345678> from 34567890 failed: ciphertext verification failed ...] [...] Fatal: repository contains errorsThis fix only applies to a very specific case where the log shows
operation successful after 1 retriesfollowed by acheck successful on second attempt, original errorthat only reportsciphertext verification failederrors in the pack file. If any other errors are reported in the pack file, then the repository still has to be considered as damaged.Now, only the check result of the last download retry is reported as intended.
restic 0.18.0
Changelog for restic 0.18.0 (2025-03-27)
The following sections list the changes in restic 0.18.0 relevant to restic users. The changes are ordered by importance.
Summary
- Sec #5291: Mitigate attack on content-defined chunking algorithm
- Fix #1843: Correctly restore long filepaths' timestamp on old Windows
- Fix #2165: Ignore disappeared backup source files
- Fix #5153: Include root tree when searching using
find --tree - Fix #5169: Prevent Windows VSS event log 8194 warnings for backup with fs snapshot
- Fix #5212: Fix duplicate data handling in
prune --max-unused - Fix #5249: Fix creation of oversized index by
repair index --read-all-packs - Fix #5259: Fix rare crash in command output
- Chg #4938: Update dependencies and require Go 1.23 or newer
- Chg #5162: Promote feature flags
- Enh #1378: Add JSON support to
checkcommand - Enh #2511: Support generating shell completions to stdout
- Enh #3697: Allow excluding online-only cloud files (e.g. OneDrive)
- Enh #4179: Add
sortoption tolscommand - Enh #4433: Change default sort order for
findoutput - Enh #4521: Add support for Microsoft Blob Storage access tiers
- Enh #4942: Add snapshot summary statistics to rewritten snapshots
- Enh #4948: Format exit errors as JSON when requested
- Enh #4983: Add SLSA provenance to GHCR container images
- Enh #5054: Enable compression for ZIP archives in
dumpcommand - Enh #5081: Add retry mechanism for loading repository config
- Enh #5089: Allow including/excluding extended file attributes during
restore - Enh #5092: Show count of deleted files and directories during
restore - Enh #5109: Make small pack size configurable for
prune - Enh #5119: Add start and end timestamps to
backupJSON output - Enh #5131: Add DragonFlyBSD support
- Enh #5137: Make
tagcommand print which snapshots were modified - Enh #5141: Provide clear error message if AZURE_ACCOUNT_NAME is not set
- Enh #5173: Add experimental S3 cold storage support
- Enh #5174: Add xattr support for NetBSD 10+
- Enh #5251: Improve retry handling for flaky
rclonebackends - Enh #52897: Make
recoverautomatically rebuild index when needed
Details
-
Security #5291: Mitigate attack on content-defined chunking algorithm
Restic uses Rabin Fingerprints for its content-defined chunker. The algorithm relies on a secret polynomial to split files into chunks.
As shown in the paper "Chunking Attacks on File Backup Services using Content-Defined Chunking" by Boris Alexeev, Colin Percival and Yan X Zhang, an attacker that can observe chunk sizes for a known file can derive the secret polynomial. Knowledge of the polynomial might in some cases allow an attacker to check whether certain large files are stored in a repository.
A practical attack is nevertheless hard as restic merges multiple chunks into opaque pack files and by default processes multiple files in parallel. This likely prevents an attacker from matching pack files to the attacker-known file and thereby prevents the attack.
Despite the low chances of a practical attack, restic now has added mitigation that randomizes how chunks are assembled into pack files. This prevents attackers from guessing which chunks are part of a pack file and thereby prevents learning the chunk sizes.
-
Bugfix #1843: Correctly restore long filepaths' timestamp on old Windows
The
restorecommand now correctly restores timestamps for files with paths longer than 256 characters on Windows versions prior to Windows 10 1607. -
Bugfix #2165: Ignore disappeared backup source files
The
backupcommand now quietly skips files that are removed between directory listing and backup, instead of printing errors like:error: lstat /some/file/name: no such file or directory -
Bugfix #5153: Include root tree when searching using
find --treeThe
restic find --treecommand did not find trees referenced byrestic snapshot --json. It now correctly includes the root tree when searching. -
Bugfix #5169: Prevent Windows VSS event log 8194 warnings for backup with fs snapshot
When running
backupwith the--use-fs-snapshotoption in Windows with admin rights, event logs likeVolume Shadow Copy Service error: Unexpected error querying for the IVssWriterCallback interface. hr = 0x80070005, Access is denied. . This is often caused by incorrect security settings in either the writer or requester process. Operation: Gathering Writer Data Context: Writer Class Id: {e8132975-6f93-4464-a53e-1050253ae220} Writer Name: System Writer Writer Instance ID: {54b151ac-d27d-4628-9cb0-2bc40959f50f}Are created several times even though the backup itself succeeds. This has now been fixed.
#5169 #5170 https://forum.restic.net/t/windows-shadow-copy-snapshot-vss-unexpected-provider-error/3674/2
-
Bugfix #5212: Fix duplicate data handling in
prune --max-unusedThe
prune --max-unused sizecommand did not correctly account for duplicate data. If a repository contained a large amount of duplicate data, this could previously result in pruning too little data. This has now been fixed.#5212 https://forum.restic.net/t/restic-not-obeying-max-unused-parameter-on-prune/8879
-
Bugfix #5249: Fix creation of oversized index by
repair index --read-all-packsSince restic 0.17.0, the new index created by
repair index --read-all-packswas written as a single large index. This significantly increased memory usage while loading the index.The index is now correctly split into multiple smaller indexes, and
repair indexnow also automatically splits oversized indexes. -
Bugfix #5259: Fix rare crash in command output
Some commands could in rare cases crash when trying to print status messages and request retries at the same time, resulting in an error like the following:
panic: runtime error: slice bounds out of range [468:156] [...] github.com/restic/restic/internal/ui/termstatus.(*lineWriter).Write(...) /restic/internal/ui/termstatus/stdio_wrapper.go:36 +0x136This has now been fixed.
-
Change #4938: Update dependencies and require Go 1.23 or newer
We have updated all dependencies. Restic now requires Go 1.23 or newer to build.
This also disables support for TLS versions older than TLS 1.2. On Windows, restic now requires at least Windows 10 or Windows Server 2016. On macOS, restic now requires at least macOS 11 Big Sur.
-
Change #5162: Promote feature flags
The
deprecate-legacy-index,deprecate-s3-legacy-layout,explicit-s3-anonymous-authandsafe-forget-keep-tagsfeatures are now stable and can no longer be disabled. The corresponding feature flags will be removed in restic 0.19.0. -
Enhancement #1378: Add JSON support to
checkcommandThe
checkcommand now supports the--jsonoption to output all statistics in JSON format. -
Enhancement #2511: Support generating shell completions to stdout
The
generatecommand now supports using-as the filename with the--[shell]-completionoption to ...
restic 0.17.3
Changelog for restic 0.17.3 (2024-11-08)
The following sections list the changes in restic 0.17.3 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #4971: Fix unusable
mounton macOS Sonoma - Fix #5003: Fix metadata errors during backup of removable disks on Windows
- Fix #5101: Do not retry load/list operation if SFTP connection is broken
- Fix #5107: Fix metadata error on Windows for backups using VSS
- Enh #5096: Allow
prune --dry-runwithout lock
Details
-
Bugfix #4971: Fix unusable
mounton macOS SonomaOn macOS Sonoma when using FUSE-T, it was not possible to access files in a mounted repository. This issue is now resolved.
-
Bugfix #5003: Fix metadata errors during backup of removable disks on Windows
Since restic 0.17.0, backing up removable disks on Windows could report errors with retrieving metadata like shown below.
error: incomplete metadata for d:\filename: get named security info failed with: Access is denied.This has now been fixed.
#5003 #5123 https://forum.restic.net/t/backing-up-a-folder-from-a-veracrypt-volume-brings-up-errors-since-restic-v17-0/8444
-
Bugfix #5101: Do not retry load/list operation if SFTP connection is broken
When using restic with the SFTP backend, backend operations that load a file or list files were retried even if the SFTP connection was broken. This has now been fixed.
#5101 https://forum.restic.net/t/restic-hanging-on-backup/8559
-
Bugfix #5107: Fix metadata error on Windows for backups using VSS
Since restic 0.17.2, when creating a backup on Windows using
--use-fs-snapshot, restic would report an error like the following:error: incomplete metadata for C:\: get EA failed while opening file handle for path \\?\GLOBALROOT\Device\HarddiskVolumeShadowCopyXX\, with: The process cannot access the file because it is being used by another process.This has now been fixed by correctly handling paths that refer to volume shadow copy snapshots.
-
Enhancement #5096: Allow
prune --dry-runwithout lockThe
prune --dry-run --no-locknow allows performing a dry-run without locking the repository. Note that if the repository is modified concurrently,prunemay return inaccurate statistics or errors.
restic 0.17.2
Changelog for restic 0.17.2 (2024-10-27)
The following sections list the changes in restic 0.17.2 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #4004: Support container-level SAS/SAT tokens for Azure backend
- Fix #5047: Resolve potential error during concurrent cache cleanup
- Fix #5050: Return error if
tagfails to lock repository - Fix #5057: Exclude irregular files from backups
- Fix #5063: Correctly
backupextended metadata when using VSS on Windows
Details
-
Bugfix #4004: Support container-level SAS/SAT tokens for Azure backend
Restic previously expected SAS/SAT tokens to be generated at the account level, which prevented tokens created at the container level from being used to initialize a repository. This caused an error when attempting to initialize a repository with container-level tokens.
Restic now supports both account-level and container-level SAS/SAT tokens for initializing a repository.
-
Bugfix #5047: Resolve potential error during concurrent cache cleanup
When multiple restic processes ran concurrently, they could compete to remove obsolete snapshots from the local backend cache, sometimes leading to a "no such file or directory" error. Restic now suppresses this error to prevent issues during cache cleanup.
-
Bugfix #5050: Return error if
tagfails to lock repositorySince restic 0.17.0, the
tagcommand did not return an error when it failed to open or lock the repository. This issue has now been fixed. -
Bugfix #5057: Exclude irregular files from backups
Since restic 0.17.1, files with the type
irregularcould mistakenly be included in snapshots, especially when backing up special file types on Windows that restic cannot process. This issue has now been fixed.Previously, this bug caused the
checkcommand to report errors like the following one:tree 12345678[...]: node "example.zip" with invalid type "irregular"To repair affected snapshots, upgrade to restic 0.17.2 and run:
restic repair snapshots --forgetThis will remove the
irregularfiles from the snapshots (creating a new snapshot ID for each of the affected snapshots).#5057 https://forum.restic.net/t/errors-found-by-check-1-invalid-type-irregular-2-ciphertext-verification-failed/8447/2
-
Bugfix #5063: Correctly
backupextended metadata when using VSS on WindowsOn Windows, when creating a backup with the
--use-fs-snapshotoption, restic read extended metadata from the original filesystem path instead of from the snapshot. This could result in errors if files were removed during the backup process.This issue has now been resolved.
restic 0.17.1
Changelog for restic 0.17.1 (2024-09-05)
The following sections list the changes in restic 0.17.1 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #2004: Correctly handle volume names in
backupcommand on Windows - Fix #4945: Include missing backup error text with
--json - Fix #4953: Correctly handle long paths on older Windows versions
- Fix #4957: Fix delayed cancellation of certain commands
- Fix #4958: Don't ignore metadata-setting errors during restore
- Fix #4969: Correctly restore timestamp for files with resource forks on macOS
- Fix #4975: Prevent
backup --stdin-from-commandfrom panicking - Fix #4980: Skip extended attribute processing on unsupported Windows volumes
- Fix #5004: Fix spurious "A Required Privilege Is Not Held by the Client" error
- Fix #5005: Fix rare failures to retry locking a repository
- Fix #5018: Improve HTTP/2 support for REST backend
- Chg #4953: Also back up files with incomplete metadata
- Enh #4795: Display progress bar for
restore --verify - Enh #4934: Automatically clear removed snapshots from cache
- Enh #4944: Print JSON-formatted errors during
restore --json - Enh #4959: Return exit code 12 for "bad password" errors
- Enh #4970: Make timeout for stuck requests customizable
Details
-
Bugfix #2004: Correctly handle volume names in
backupcommand on WindowsOn Windows, when the specified backup target only included the volume name without a trailing slash, for example,
C:, then restoring the resulting snapshot would result in an error. Note that usingC:\as backup target worked correctly.Specifying volume names is now handled correctly. To restore snapshots created before this bugfix, use the : syntax. For example, to restore a snapshot with ID
12345678that backed upC:, use the following command:restic restore 12345678:/C/C:./ --target output/folder -
Bugfix #4945: Include missing backup error text with
--jsonPreviously, when running a backup with the
--jsonoption, restic failed to include the actual error message in the output, resulting in"error": {}being displayed.This has now been fixed, and restic now includes the error text in JSON output.
-
Bugfix #4953: Correctly handle long paths on older Windows versions
On older Windows versions, like Windows Server 2012, restic 0.17.0 failed to back up files with long paths. This problem has now been resolved.
-
Bugfix #4957: Fix delayed cancellation of certain commands
Since restic 0.17.0, some commands did not immediately respond to cancellation via Ctrl-C (SIGINT) and continued running for a short period. The most affected commands were
diff,find,ls,statsandrewrite. This is now resolved. -
Bugfix #4958: Don't ignore metadata-setting errors during restore
Previously, restic used to ignore errors when setting timestamps, attributes, or file modes during a restore. It now reports those errors, except for permission related errors when running without root privileges.
-
Bugfix #4969: Correctly restore timestamp for files with resource forks on macOS
On macOS, timestamps were not restored for files with resource forks. This has now been fixed.
-
Bugfix #4975: Prevent
backup --stdin-from-commandfrom panickingRestic would previously crash if
--stdin-from-commandwas specified without providing a command. This issue has now been fixed. -
Bugfix #4980: Skip extended attribute processing on unsupported Windows volumes
With restic 0.17.0, backups of certain Windows paths, such as network drives, failed due to errors while fetching extended attributes.
Restic now skips extended attribute processing for volumes where they are not supported.
-
Bugfix #5004: Fix spurious "A Required Privilege Is Not Held by the Client" error
On Windows, creating a backup could sometimes trigger the following error:
error: nodeFromFileInfo [...]: get named security info failed with: a required privilege is not held by the client.This has now been fixed.
-
Bugfix #5005: Fix rare failures to retry locking a repository
Restic 0.17.0 could in rare cases fail to retry locking a repository if one of the lock files failed to load, resulting in the error:
unable to create lock in backend: circuit breaker open for file <lock/1234567890>This issue has now been addressed. The error handling now properly retries the locking operation. In addition, restic waits a few seconds between locking retries to increase chances of successful locking.
-
Bugfix #5018: Improve HTTP/2 support for REST backend
If
rest-servertried to gracefully shut down an HTTP/2 connection still in use by the client, it could result in the following error:http2: Transport: cannot retry err [http2: Transport received Server's graceful shutdown GOAWAY] after Request.Body was written; define Request.GetBody to avoid this errorThis issue has now been resolved.
#5018 https://forum.restic.net/t/receiving-http2-goaway-messages-with-windows-restic-v0-17-0/8367
-
Change #4953: Also back up files with incomplete metadata
If restic failed to read extended metadata for a file or folder during a backup, then the file or folder was not included in the resulting snapshot. Instead, a warning message was printed along with returning exit code 3 once the backup was finished.
Now, restic also includes items for which the extended metadata could not be read in a snapshot. The warning message has been updated to:
incomplete metadata for /path/to/file: <details about error> -
Enhancement #4795: Display progress bar for
restore --verifyWhen the
restorecommand is run with--verify, it now displays a progress bar while the verification step is running. The progress bar is not shown when the--jsonflag is specified. -
Enhancement #4934: Automatically clear removed snapshots from cache
Previously, restic only removed snapshots from the cache on the host where the
forgetcommand was executed. On other hosts that use the same repository, the old snapshots remained in the cache.Restic now automatically clears old snapshots from the local cache of the current host.
-
Enhancement #4944: Print JSON-formatted errors during
restore --jsonRestic used to print any
restoreerrors directly to the console as freeform text messages, even when using the--jsonoption.Now, when
--jsonis specified, restic prints them as JSON formatted messages. -
Enhancement #4959: Return exit code 12 for "bad password" errors
Restic now returns exit code 12 when it cannot open the repository due to an incorrect password.
-
Enhancement #4970: Make timeout for stuck requests customizable
Restic monitors connections to the backend to detect stuck requests. If a request does not return any data within five minutes, restic assumes the request is stuck and retries it. However, for large repositories this timeout might be insufficient to collect a list of all files, causing the following error:
`List(data) returned error,...
restic 0.17.0
Changelog for restic 0.17.0 (2024-07-26)
The following sections list the changes in restic 0.17.0 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #3600: Handle unreadable xattrs in folders above
backupsource - Fix #4209: Fix slow SFTP upload performance
- Fix #4503: Correct hardlink handling in
statscommand - Fix #4568: Prevent
forget --keep-tags <invalid>from deleting all snapshots - Fix #4615: Make
findnot sometimes ignore directories - Fix #4656: Properly report ID of newly added keys
- Fix #4703: Shutdown cleanly when receiving SIGTERM
- Fix #4709: Correct
--no-lockhandling oflsandtagcommands - Fix #4760: Fix possible error on concurrent cache cleanup
- Fix #4850: Handle UTF-16 password files in
keycommand correctly - Fix #4902: Update snapshot summary on
rewrite - Chg #956: Return exit code 10 and 11 for non-existing and locked repository
- Chg #4540: Require at least ARMv6 for ARM binaries
- Chg #4602: Deprecate legacy index format and
s3legacyrepository layout - Chg #4627: Redesign backend error handling to improve reliability
- Chg #4707: Disable S3 anonymous authentication by default
- Chg #4744: Include full key ID in JSON output of
key list - Enh #662: Optionally skip snapshot creation if nothing changed
- Enh #693: Include snapshot size in
snapshotsoutput - Enh #805: Add bitrot detection to
diffcommand - Enh #828: Improve features of the
repair packscommand - Enh #1786: Support repositories with empty password
- Enh #2348: Add
--deleteoption torestorecommand - Enh #3067: Add extended options to configure Windows Shadow Copy Service
- Enh #3406: Improve
dumpperformance for large files - Enh #3806: Optimize and make
prunecommand resumable - Enh #4006: (alpha) Store deviceID only for hardlinks
- Enh #4048: Add support for FUSE-T with
mounton macOS - Enh #4251: Support reading backup from a command's standard output
- Enh #4287: Support connection to rest-server using unix socket
- Enh #4354: Significantly reduce
prunememory usage - Enh #4437: Make
checkcommand create non-existent cache directory - Enh #4472: Support AWS Assume Role for S3 backend
- Enh #4547: Add
--jsonoption toversioncommand - Enh #4549: Add
--ncduoption tolscommand - Enh #4573: Support rewriting host and time metadata in snapshots
- Enh #4583: Ignore
s3.storage-classarchive tiers for metadata - Enh #4590: Speed up
mountcommand's error detection - Enh #4601: Add support for feature flags
- Enh #4611: Back up more file metadata on Windows
- Enh #4664: Make
lsusemessage_typefield in JSON output - Enh #4676: Make
keycommand's actions separate sub-commands - Enh #4678: Add
--targetoption to thedumpcommand - Enh #4708: Back up and restore SecurityDescriptors on Windows
- Enh #4733: Allow specifying
--hostvia environment variable - Enh #4737: Include snapshot ID in
reasonfield offorgetJSON output - Enh #4764: Support forgetting all snapshots
- Enh #4768: Allow specifying custom User-Agent for outgoing requests
- Enh #4781: Add
restoreoptions to read include/exclude patterns from files - Enh #4807: Support Extended Attributes on Windows NTFS
- Enh #4817: Make overwrite behavior of
restorecustomizable - Enh #4839: Add dry-run support to
restorecommand
Details
-
Bugfix #3600: Handle unreadable xattrs in folders above
backupsourceWhen backup sources are specified using absolute paths,
backupalso includes information about the parent folders of the backup sources in the snapshot.If the extended attributes for some of these folders could not be read due to missing permissions, this caused the backup to fail. This has now been fixed.
#3600 #4668 https://forum.restic.net/t/parent-directories-above-the-snapshot-source-path-fatal-error-permission-denied/7216
-
Bugfix #4209: Fix slow SFTP upload performance
Since restic 0.12.1, the upload speed of the sftp backend to a remote server has regressed significantly. This has now been fixed.
-
Bugfix #4503: Correct hardlink handling in
statscommandIf files on different devices had the same inode ID, the
statscommand did not correctly calculate the snapshot size. This has now been fixed.#4503 #4006 https://forum.restic.net/t/possible-bug-in-stats/6461/8
-
Bugfix #4568: Prevent
forget --keep-tags <invalid>from deleting all snapshotsRunning
forget --keep-tags <invalid>, where<invalid>is a tag that does not exist in the repository, would remove all snapshots. This is especially problematic if the tag name contains a typo.The
forgetcommand now fails with an error if all snapshots in a snapshot group would be deleted. This prevents the above example from deleting all snapshots.It is possible to temporarily disable the new check by setting the environment variable
RESTIC_FEATURES=safe-forget-keep-tags=false. Note that this feature flag will be removed in the next minor restic version. -
Bugfix #4615: Make
findnot sometimes ignore directoriesIn some cases, the
findcommand ignored empty or moved directories. This has now been fixed. -
Bugfix #4656: Properly report ID of newly added keys
restic key addnow reports the ID of the newly added key. This simplifies selecting a specific key using the--key-hint keyoption. -
Bugfix #4703: Shutdown cleanly when receiving SIGTERM
Previously, when restic received the SIGTERM signal it would terminate immediately, skipping cleanup and potentially causing issues like stale locks being left behind. This primarily effected containerized restic invocations that use SIGTERM, but could also be triggered via a simple
killall restic.This has now been fixed, such that restic shuts down cleanly when receiving the SIGTERM signal.
-
Bugfix #4709: Correct
--no-lockhandling oflsandtagcommandsThe
lscommand never locked the repository. This has now been fixed, with the old behavior still being supported usingls --no-lock. The latter invocation also works with older restic versions.The
tagcommand erroneously accepted the--no-lockcommand. This command now always requires an exclusive lock. -
Bugfix #4760: Fix possible error on concurrent cache cleanup
If multiple restic processes concurrently cleaned up no longer existing files from the cache, this could cause some of the processes to fail with an
no such file or directoryerror. This has now been fixed.
restic 0.16.5
Changelog for restic 0.16.5 (2024-07-01)
The following sections list the changes in restic 0.16.5 relevant to restic users. The changes are ordered by importance.
Summary
Details
-
Enhancement #4799: Add option to force use of Azure CLI credential
A new environment variable
AZURE_FORCE_CLI_CREDENTIAL=trueallows forcing the use of Azure CLI credential, ignoring other credentials like managed identity. -
Enhancement #4873: Update dependencies
A few potentially vulnerable dependencies were updated.
restic 0.16.4
This release works around and improves detection of a bug in the compression library used by restic. The resulting issue only happens when using restic 0.16.3 and the max compression level (the default auto and off compression levels are not affected), and when the source files being backed up have specific data in them to trigger the bug. If you use max compression, you can use restic check --read-data to make sure you're not affected.
restic is distributed as a standalone binary: download the correct file for your operating system and architecture, extract the file and just run it. If you run into any issues, please report them at the GitHub issue tracker or visit the forum. If you already have restic >= 0.9.4, you can use restic self-update to get the latest version in a secure way.
The binaries released with each restic version are reproducible, which means that you can reproduce a byte identical version from the source code for that release. Instructions on how to do that in the Developer Documentation.
Changelog for restic 0.16.4 (2024-02-04)
The following sections list the changes in restic 0.16.4 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #4677: Downgrade zstd library to fix rare data corruption at max. compression
- Enh #4529: Add extra verification of data integrity before upload
Details
-
Bugfix #4677: Downgrade zstd library to fix rare data corruption at max. compression
In restic 0.16.3, backups where the compression level was set to
max(using--compression max) could in rare and very specific circumstances result in data corruption due to a bug in the library used for compressing data. Restic 0.16.1 and 0.16.2 were not affected.Restic now uses the previous version of the library used to compress data, the same version used by restic 0.16.2. Please note that the
autocompression level (which restic uses by default) was never affected, and even if you usedmaxcompression, chances of being affected by this issue are small.To check a repository for any corruption, run
restic check --read-data. This will download and verify the whole repository and can be used at any time to completely verify the integrity of a repository. If thecheckcommand detects anomalies, follow the suggested steps. -
Enhancement #4529: Add extra verification of data integrity before upload
Hardware issues, or a bug in restic or its dependencies, could previously cause corruption in the files restic created and stored in the repository. Detecting such corruption previously required explicitly running the
check --read-dataorcheck --read-data-subsetcommands.To further ensure data integrity, even in the case of hardware issues or software bugs, restic now performs additional verification of the files about to be uploaded to the repository.
These extra checks will increase CPU usage during backups. They can therefore, if absolutely necessary, be disabled using the
--no-extra-verifyglobal option. Please note that this should be combined with more active checking using the previously mentioned check commands.
restic 0.16.3
This release fixes a couple of bugs on Windows and in the restore command. It also works around an unlikely yet possible situation with rclone which could potentially result in data loss.
restic is distributed as a standalone binary: download the correct file for your operating system and architecture, extract the file and just run it. If you run into any issues, please report them at the GitHub issue tracker or visit the forum. If you already have restic >= 0.9.4, you can use restic self-update to get the latest version in a secure way.
The binaries released with each restic version are reproducible, which means that you can reproduce a byte identical version from the source code for that release. Instructions on how to do that in the Developer Documentation.
Changelog for restic 0.16.3 (2024-01-14)
The following sections list the changes in restic 0.16.3 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #4560: Improve errors for irregular files on Windows
- Fix #4574: Support backup of deduplicated files on Windows again
- Fix #4612: Improve error handling for
rclonebackend - Fix #4624: Correct
restoreprogress information if an error occurs - Fix #4626: Improve reliability of restoring large files
Details
-
Bugfix #4560: Improve errors for irregular files on Windows
Since Go 1.21, most filesystem reparse points on Windows are considered to be irregular files. This caused restic to show an
error: invalid node type ""error message for those files.This error message has now been improved and includes the relevant file path:
error: nodeFromFileInfo path/to/file: unsupported file type "irregular". As irregular files are not required to behave like regular files, it is not possible to provide a generic way to back up those files.#4560 #4620 https://forum.restic.net/t/windows-backup-error-invalid-node-type/6875
-
Bugfix #4574: Support backup of deduplicated files on Windows again
With the official release builds of restic 0.16.1 and 0.16.2, it was not possible to back up files that were deduplicated by the corresponding Windows Server feature. This also applied to restic versions built using Go 1.21.0-1.21.4.
The Go version used to build restic has now been updated to fix this.
-
Bugfix #4612: Improve error handling for
rclonebackendSince restic 0.16.0, if rclone encountered an error while listing files, this could in rare circumstances cause restic to assume that there are no files. Although unlikely, this situation could result in data loss if it were to happen right when the
prunecommand is listing existing snapshots.Error handling has now been improved to detect and work around this case.
-
Bugfix #4624: Correct
restoreprogress information if an error occursIf an error occurred while restoring a snapshot, this could cause the
restoreprogress bar to show incorrect information. In addition, if a data file could not be loaded completely, then errors would also be reported for some already restored files.Error reporting of the
restorecommand has now been made more accurate.#4624 https://forum.restic.net/t/errors-restoring-with-restic-on-windows-server-s3/6943
-
Bugfix #4626: Improve reliability of restoring large files
In some cases restic failed to restore large files that frequently contain the same file chunk. In combination with certain backends, this could result in network connection timeouts that caused incomplete restores.
Restic now includes special handling for such file chunks to ensure reliable restores.
#4626 https://forum.restic.net/t/errors-restoring-with-restic-on-windows-server-s3/6943
restic 0.16.2
restic 0.16.2 restores support for ARMv5 which was accidentally not included in the previous one, and also fixes missing documentation on Read the Docs.
restic is distributed as a standalone binary: download the correct file for your operating system and architecture, extract the file and just run it. If you run into any issues, please report them at the GitHub issue tracker or visit the forum. If you already have restic >= 0.9.4, you can use restic self-update to get the latest version in a secure way.
The binaries released with each restic version are reproducible, which means that you can reproduce a byte identical version from the source code for that release. Instructions on how to do that in the Developer Documentation.
Changelog for restic 0.16.2 (2023-10-29)
The following sections list the changes in restic 0.16.2 relevant to restic users. The changes are ordered by importance.
Summary
- Fix #4540: Restore ARMv5 support for ARM binaries
- Fix #4545: Repair documentation build on Read the Docs
Details
-
Bugfix #4540: Restore ARMv5 support for ARM binaries
The official release binaries for restic 0.16.1 were accidentally built to require ARMv7. The build process is now updated to restore support for ARMv5.
Please note that restic 0.17.0 will drop support for ARMv5 and require at least ARMv6.
-
Bugfix #4545: Repair documentation build on Read the Docs
For restic 0.16.1, no documentation was available at https://restic.readthedocs.io/ .
The documentation build process is now updated to work again.