Repository navigation
Releases
v0.36.0
DittoFS v0.36.0
A modular virtual filesystem that decouples file interfaces from storage backends.
What's new
0.36.0 is a fix release, mostly for SMB. Before upgrading, check any SMB adapter config you have set by hand: the SMB adapter now refuses an encryption object it cannot apply, both when the config is stored and at boot. A stored row that 0.35.0 accepted can keep the SMB adapter down after the upgrade. See Changed below. No on-disk format changes in this release.
Changed
-
The SMB adapter refuses an
encryptionsetting that it would otherwise ignore.encryptionmust be an object holding onlyencryption_modeandallowed_ciphers, andencryption_modemust be a string. Before, a non-objectencryption, a non-string mode or a misspelt key such asmodewas read as no mode at all. The adapter stayed onpreferredanddfsctlreported success, so an operator who asked forrequiredstill served unencrypted SMB 2.x. -
allowed_cipherscan no longer be set; the SMB cipher set is fixed. The key was accepted with any value and never applied: NEGOTIATE kept offering every cipher whiledfsctl adapter edit smb --configreported success. A non-empty value is now refused withallowed_ciphers cannot be configured: the SMB cipher set is fixed. An absent key,nulland[]are still accepted. Negotiation itself is unchanged.Both checks run when the config is stored and when the server boots. A stored row that fails them leaves the SMB adapter down with
Adapter failed to start; continuing without it, while the server, the API and NFS keep running. To recover, store the config again without the offending key, for exampledfsctl adapter edit smb --config '{"encryption":{"encryption_mode":"required"}}'. The adapter then starts without a server restart. Adapters seeded by default carry no config and are not affected.
Security
-
encryption.encryption_modein the SMB adapter config now takes effect.dfs startreadbind_addressandsigningout of the adapter config but skippedencryption_mode, so the process stayed on thepreferreddefault andrequiredcould not be selected from the adapter config at all. -
SMB refuses a security descriptor holding an ACE type it cannot store, instead of dropping that ACE. Object ACEs, callback ACEs, mandatory labels and other types the server does not model were skipped, the rest of the DACL was stored, and the request succeeded. A skipped deny granted what the client had denied; a skipped allow silently withdrew a grant. SET_INFO now answers
STATUS_INVALID_PARAMETERand leaves the stored ACL as it was. A CREATE whose security descriptor fails to parse or validate is answered the same way and creates nothing, where it used to create the file without the descriptor. A SACL is parsed whenever one is present, so a SACL carrying anything other than audit or alarm ACEs, a mandatory label for instance, now fails the request too, even when the client did not ask for the SACL to be set. -
SMB TREE_CONNECT is refused when the user's share permission cannot be resolved. A store error used to fall back to the share's default permission, which can be above a user's explicit grant, or above an explicit
none. Since a tree keeps the permission it was connected with until a control-plane change re-checks it, the fallback outlived the failure. TREE_CONNECT now answersSTATUS_ACCESS_DENIED, as NFS already refuses, and the client can retry once the store answers. A failed lookup of a principal's SID grants refuses as well.
Fixed
-
Truncating two files down to the same content no longer fails forever on the second one. The second truncate was refused on every retry, as
NFS4ERR_DELAYover NFS andSTATUS_INSUFFICIENT_RESOURCESover SMB. Truncate gave each file an identifier derived from its remaining content, which the Badger store kept unique, although nothing read it any more. That identifier and its index are now removed. The on-disk format stays compatible: stores and backups written by older versions still open, a file record drops the old field the next time it is written, and the leftover index keys are deleted when a store is first opened or right after a restore, with a log line (badger: dropped the retired ObjectID index). On a store with a million such keys this takes about 12 seconds. Rolling back to an older version can hit the bug again. -
Opening an existing file for append over NFSv3 no longer empties it. A CREATE in UNCHECKED mode that named an existing file truncated it to zero and applied the request's mode and ownership to it. Linux clients send that CREATE for
open(O_CREAT), so on some kernelsecho x >> filelost the file's contents and wrote at offset 0. Now, as on the Linux kernel server, only a size the request carries is applied; without one the file is left alone. On a read-only share such a CREATE fails withNFS3ERR_ROFS, so a client opening the file for writing learns at open rather than at its first write. An existing directory or symlink named without a size is answered with its own handle and attributes. NFSv4 was not affected. -
A Windows client can open an encrypted share when the session does not require encryption. In
preferredmode the server encrypted the TREE_CONNECT response for a share that requires encryption, but the client learns that from the flag inside that response and was not decrypting yet. Windows retried and dropped the connection. TREE_CONNECT is now sent in the clear in that case; later commands on the tree are still encrypted, and a session that already requires encryption still encrypts TREE_CONNECT. -
SMB Kerberos accepts a ticket whose addresses are NetBIOS names only. Windows often lists the workstation's NetBIOS name and no IP in a ticket's addresses, and comparing those to the connection's IP failed session setup with
KRB_AP_ERR_BADADDR. Non-IP addresses no longer constrain the peer. A ticket that lists any IP still has to match the connection's IP. Verify failures now logclient_addressandticket_addresses. NFS is unchanged. -
A whole-share warm no longer holds memory in proportion to the share while its progress reporting is stalled. Each finished block queued one progress event until the reporter caught up; pending progress is now two counters. Progress calls may therefore jump by more than one block, and a successful warm still ends with done equal to total.
-
The operator now shows when the DittoFS API refuses its own account. Applying
spec.snapshotPoliciesand turning on the NFS portmapper both need the admin role, so the operator's account (roleoperator) got 403 on each, and that showed only in a verbose log line. A newAuthorizedcondition on the DittoServer is False and names each refused call. It turns True once every call gets its answer, and a transient API error leaves it as it was. The NFS adapter's Service now publishes port 111 only once the operator has seen the portmapper on, so it no longer points at a port nothing listens on. The operator still cannot apply these settings without the admin role.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.36.0/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.36.0/dfs_0.36.0_amd64.pkg.tar.zst
sudo pacman -U dfs_0.36.0_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.36.0Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.36.0/dfs_0.36.0_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.36.0/dfsctl_0.36.0_Linux_x86_64.tar.gz | tar xzChangelog
Features
- 9e71876: feat(harness): JUnit per cell, generated from summary.json (#3168) (@llucian-iws)
- 3c277d2: feat(harness): NFS mounts in the fixture (#3183) (@llucian-iws)
- 9008c6f: feat(harness): S3 profiles: the bucket, its write-and-read check, and the localstack need (#3181) (@llucian-iws)
- d21f184: feat(harness): bring dt in line with the settled step 2 gaps (#3184) (@llucian-iws)
- 6928f60: feat(harness): commands in images, with templates (#3166) (@llucian-iws)
- e2a80ec: feat(harness): dt doctor, and dt list says whether each suite can run here (#3123) (@llucian-iws)
- 7efc6e3: feat(harness): dt list and dt plan from the suite registry (#3088) (@llucian-iws)
- c95b66d: feat(harness): dt report, and job summaries from ...
v0.35.0
DittoFS v0.35.0
A modular virtual filesystem that decouples file interfaces from storage backends.
What's new
The main change in 0.35.0 is that BadgerDB is now the only metadata store: the PostgreSQL, SQLite and legacy memory metadata store backends have been removed. Read the migration guide below before upgrading a server that uses any of them.
Removed
- Breaking: the PostgreSQL, SQLite and legacy
memorymetadata stores are gone; BadgerDB is the only metadata store. Thepostgres,sqliteandmemorymetadata store types no longer exist. A server whose configuration still names one refuses to start, withunsupported metadata store type, and changes nothing. There is no conversion tool: a share on a SQL store moves by copying its files to a share on a new instance. Amemorystore is removed before the upgrade and recreated as a Badger store kept in RAM, withdfsctl store metadata add --name <store> --in-memory. See the metadata store migration guide before upgrading. The control-plane database (database:, SQLite or PostgreSQL) is not affected.
Fixed
-
A write refused because the local store is full now reaches the client as "no space", not as an I/O error. SMB returns
STATUS_DISK_FULL, NFSv3NFS3ERR_NOSPCand NFSv4NFS4ERR_NOSPC, where all three used to return their generic I/O error. -
A request that arrives with no deadline gets 30 seconds at the block store. A cold read from a remote that stopped answering, or a write held while the local store was at capacity, used to wait out their own 60-second budgets, and the backpressure budget restarted whenever the remote drained a little.
-
Deleting a file no longer waits for a flush pass to finish uploading. The delete itself was already durable; only the best-effort reclaim at its end waited, and with the remote down it waited for upload slots that never freed. An SMB delete-on-close was measured held for 68 seconds, past the point where clients give up on the connection.
-
A write that lands while a CLONE replaces its destination is no longer read back as zeros. A write outside the cloned range could commit between the clone's metadata transaction and its journal discard and be lost, and a background carve or warm could republish the old content over the clone. CLONE now holds the source and destination exclusively for the whole replacement, and ordinary writes, carves and warms take shared ownership.
-
That ownership is no longer held across a clone's upload or a whole-share warm. A CLONE drains its source before taking the exclusive hold, and a warm holds one payload at a time, so reads and writes on other files no longer queue behind either.
-
Directory removal and rename are ordered against creates inside the directory. A directory could be removed after another operation had created or moved a child into it, a create could commit after its parent was gone, and a delayed removal could delete a name that a rename had already replaced. Creates of distinct names in one directory still run concurrently.
-
New files created over SMB get the user's own primary GID, as they do over NFS. A user with a GID and no group memberships used to write files owned by group 1000, and a user with groups got their first group's GID instead of their own.
-
dfsctl store block add --configno longer drops the flags given beside it.--compression, every--encryption-*flag,--parallel-uploadsand the S3 field flags were silently discarded, so a store asked for compression or encryption was created without it. Flags are now merged into the JSON, and a flag that contradicts a key the JSON already sets is refused.--type memorynow sends the compression, encryption and upload settings as well. -
An S3 endpoint refused by the private-address guard fails at once. The SDK retried the refused dial up to ten times with backoff, turning a definite refusal into a 30-second timeout.
-
The startup warning for a missing config file names
dfs init, notdittofs init.
Performance
- Sequential reads of a file that is already local no longer re-download it. Prefetch ignored what the journal already held, so each warm 32 MiB pass fetched 24 MiB from the remote again; it now fetches only the ranges that are not resident. Demand fetches and integrity checks are unchanged.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.35.0/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.35.0/dfs_0.35.0_amd64.pkg.tar.zst
sudo pacman -U dfs_0.35.0_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.35.0Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.35.0/dfs_0.35.0_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.35.0/dfsctl_0.35.0_Linux_x86_64.tar.gz | tar xzChangelog
Bug Fixes
- f218342: fix(adapter): give a request without a deadline 30 s at the block store (#2966) (@llucian-iws)
- 8d18712: fix(adapter): refuse a write to a full local store as no space, not I/O (#2961) (@llucian-iws)
- 468a373: fix(cli): the startup warning names the dfs binary, not dittofs (#3025) (@viquitorreis)
- 4dae556: fix(dfsctl): merge flags given with store block add --config (#2974) (@llucian-iws)
- f676b91: fix(docs): add client prerequisites and fix brew and SMB mount steps in Getting Started (@viquitorreis)
- 246e6c9: fix(docs): write the NFS test file to the right mount path on macOS (@viquitorreis)
- 68946fe: fix(engine): don't hold payload scopes across a clone's upload or a whole-share warm (#3030) (@llucian-iws)
- 230508c: fix(journal): don't make Delete wait for a flush pass to reclaim (#2965) (@llucian-iws)
- 2142145: fix(metadata): synchronize directory lifetime with namespace mutations (#2956) (@badrkamel)
- 2a3d7f7: fix(nfs): preserve concurrent writes across clone replacement (#2958) (@badrkamel)
- 1507bdd: fix(s3): do not retry a dial the endpoint guard refused (@viquitorreis)
- e10b208: fix(s3): merge pull request #3023 from viquitorreis/fix/s3-conformance-private-endpoint-2925 (@viquitorreis)
- 2150727: fix(smb): give new files the user's own primary GID, as NFS does (#2963) (@llucian-iws)
- d0efaa7: fix(test): grade WPTS timeout and aborted outcomes as failures (#2954) (@badrkamel)
Performance Improvements
- 3e49cea: perf(engine): skip prefetch for resident journal ranges (#2957) (@badrkamel)
Others
- 2910947: Delete test/scenarios/plan directory (#2933) (@rif)
- d5f0c87: chore(deps): bump sigs.k8s.io/controller-runtime (#3015) (@dependabot[bot])
- 31757d4: chore(deps): bump the go-dependencies group with 3 updates (#3016) (@dependabot[bot])
- 8c75faf: chore(deps): bump the go-dependencies group with 6 updates (#2935) (@dependabot[bot])
- 77062d9: chore(metadata): remove the SQL and in-memory metadata stores (#2914) (@marmos91)
- aea1948: chore(release): v0.35.0 (#3047) (@marmos91)
- f854d99: ci(conformance): create the S3 bucket for pjdfstest/pynfs -s3 cells (@didimelli)
- 7f025cc: ci(dfsbench): merge pull request #2976 from viquitorreis/ci/dfsbench-smoke-dit-29 (@viquitorreis)
- ca04709: ci(dfsbench): parse the mailbox job files in the smoke job (@viquitorreis)
- 0b9ec95: ci(dfsbench): run the dfsbench smoke benchmark on pull requests (@viquitorreis)
- e6283ac: ci(e2e): run the Kerberos tests on PRs that change GSS dispatch, SECINFO or SMB identity mapping (#3045) (@llucian-iws)
- ea9a308: ci(integration): merge pull request #3024 from viquitorreis/ci/portmap-test-count-2998 (@viquitorreis)
- 5027238: ci(integration): run the portmap system test with -count=1 so its result is ne...
v0.34.1
DittoFS v0.34.1
A modular virtual filesystem that decouples file interfaces from storage backends.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.34.1/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.34.1/dfs_0.34.1_amd64.pkg.tar.zst
sudo pacman -U dfs_0.34.1_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.34.1Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.34.1/dfs_0.34.1_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.34.1/dfsctl_0.34.1_Linux_x86_64.tar.gz | tar xzChangelog
Bug Fixes
- 9a84f66: fix(ci): pass the released tag to the docs refresh workflow (#2929) (@viquitorreis)
- ce32f4e: fix(gc): free a dead block only through the remote that holds it (#2909) (#2919) (@llucian-iws)
- 18d85ae: fix(journal): keep GC's claim on a segment its repack retired (#2908) (#2918) (@llucian-iws)
- 49a583e: fix(journal): refuse incomplete sealed segment recovery (#2877) (@badrkamel)
- 814d4b5: fix(nfs): preserve destination tails on range clones (#2878) (@badrkamel)
- b207d5a: fix(smb): record ShareAccess on share-root opens (#2915) (@llucian-iws)
- 4addba0: fix(store): refuse an upgrade before renaming share columns (#2916) (@marmos91)
Others
- e65d1a5: Merge pull request #2932 from salamotoarelor/dev/isolated-scenarios (@rif)
- 3099a58: chore(deps): bump the go-dependencies group with 6 updates (#2905) (@dependabot[bot])
- 6146bfb: chore(deps): bump the operator-dependencies group (#2904) (@dependabot[bot])
- 1d85540: chore(deps): bump the operator-dependencies group (#2934) (@dependabot[bot])
- 065cdb7: chore(graphify): prune tooling dirs and document the hook (#2930) (@marmos91)
- ebd28f1: chore(release): v0.34.1 (@marmos91)
- d562e66: docs(smb): fix Windows non-default port mount instructions (#2912) (@rif)
- ed43ff6: refactor(common): delete the callerless CopyPayload helper (#2902) (@marmos91)
- bf6f371: test(scenarios): system tests, an isolated scenario runner and 38 scenarios (@rif)
Full Changelog: v0.34.0...v0.34.1
Verifying Checksums
sha256sum -c checksums.txtVerifying Signatures
Release checksums are signed with Sigstore cosign using keyless OIDC:
cosign verify-blob \
--bundle checksums.txt.sigstore.json \
--certificate-oidc-issuer https://token.actions.githubusercontent.com \
--certificate-identity-regexp '^https://github\.com/marmos91/dittofs/\.github/workflows/release\.yml@refs/tags/' \
checksums.txtDocker Images
Multi-architecture images are available for linux/amd64 and linux/arm64:
marmos91c/dittofs:v0.34.1- This specific versionmarmos91c/dittofs:latest- Latest stable releasemarmos91c/dittofs:v0- Latest v0.x.xmarmos91c/dittofs:v0.34- Latest v0.34.x
v0.34.0
DittoFS v0.34.0
A modular virtual filesystem that decouples file interfaces from storage backends.
Upgrade notes
The metadata store's on-disk format moves to version 2, and the move is one-way.
Once a v0.34.0 server opens a Badger metadata store, v0.33.0 and earlier will refuse to open it and report a future-format error. Nothing is lost and nothing is corrupted, but there is no downgrade path short of restoring a backup taken before the upgrade. Take one first if you need the option.
Existing stores migrate themselves. A file's chunk list is now held in segments rather than one value, reads accept both shapes, and the old whole-list key is retired the first time the file is written. There is no migration step to run.
Why it changed: a chunk list past a few thousand refs crossed Badger's large-value threshold, and every later commit appended a fresh copy to a log that its own reclamation pass never learned to collect. One field store reached 245 GiB of log against 0.15 GiB of live data.
Two operations that used to succeed now fail
Renaming a directory into its own subtree returns EINVAL. It used to be accepted, and it built a cycle of directories that no path reached, that the link count reported as alive, and that nothing ever collected. Any client with write permission on two directories could create one. Cycles created by an earlier build are still there and still uncollected.
Setting extended attributes past 256 KiB per file returns NFS4ERR_XATTR2BIG, or STATUS_EA_TOO_LARGE over SMB. The limit is on the encoded total for one file, not on each value, and it is all-or-nothing: nothing is stored when it trips. Per-value limits were already enforced at 64 KiB. For scale, ext4 bounds a file's whole xattr set to a single filesystem block, typically 4 KiB.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.34.0/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.34.0/dfs_0.34.0_amd64.pkg.tar.zst
sudo pacman -U dfs_0.34.0_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.34.0Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.34.0/dfs_0.34.0_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.34.0/dfsctl_0.34.0_Linux_x86_64.tar.gz | tar xzChangelog
Features
Bug Fixes
- 927bef7: fix(bench): collect parseable repeated blockstore samples (#2862) (@badrkamel)
- 60c9a6e: fix(engine): read the syncer's wiring under the lock that publishes it (#2856) (@marmos91)
- b8d0934: fix(hooks): check module tidiness without rewriting files (#2847) (@badrkamel)
- 79561cd: fix(hooks): test operator packages own module (#2848) (@badrkamel)
- 142c9f8: fix(journal): carry a retired segment's markers forward before unlinking it (#2854) (@marmos91)
- a116793: fix(journal): wire the local-tier eviction and backpressure metrics (#2872) (@marmos91)
- d122bc4: fix(metadata): bound badger conflict retries by deadline and randomise the backoff (#2889) (@marmos91)
- 654fdbd: fix(metadata): bound the total extended-attribute bytes a file may carry (#2888) (@marmos91)
- 226532e: fix(metadata): refuse a rename that moves a directory under itself (#2887) (@marmos91)
- d8591b2: fix(metadata): segment the badger chunk manifest so it stays out of the value log (#2882) (@marmos91)
- 1cfaa29: fix(smb): enumerate every metadata directory page (#2879) (@badrkamel)
- 995589b: fix(test): cap conformance grader exit statuses (#2876) (@badrkamel)
- 0830990: fix(test): reject incomplete POSIX TAP results (#2874) (@badrkamel)
- 526dba4: fix(test): size the hot-file retry test inside the budget it asserts (#2890) (@marmos91)
- 9f00199: fix: two flakes at the cause, two Wave 1 deletion gaps (#2859) (@marmos91)
Performance Improvements
- 81a5716: perf(badger): avoid repeated sparse chunk index scans (#2875) (@badrkamel)
Others
- 3e47bf6: build: compile complete CLI entrypoint packages (#2868) (@badrkamel)
- 4f878b2: chore(planning): drop spent audit and perf docs (@marmos91)
- 7cb442e: chore(planning): drop spent wave and step docs, make MEMORY.md an index (#2830) (@marmos91)
- 71cf4e1: chore(release): v0.34.0 (@marmos91)
- 13cde69: ci: build conformance binaries in-job instead of via artifacts (#2844) (@badrkamel)
- 9281dcd: ci: deduplicate Nix cache publication across test matrices (#2861) (@badrkamel)
- 9ad1d8a: ci: key the operator module cache on its own go.mod (#2835) (@badrkamel)
- 8eae797: ci: prevent vendor helper from seeding an empty Go cache (#2864) (@badrkamel)
- b1ec969: ci: run tests when SQL migrations change (#2836) (@badrkamel)
- 8da8fd3: ci: run tests when bench workloads change (#2845) (@badrkamel)
- d650b7d: ci: stagger periodic conformance suite starts (#2866) (@badrkamel)
- f72278b: ci: start conformance services only when required (#2834) (@badrkamel)
- faca5a1: ci: use the codecov files input (#2837) (@badrkamel)
- 0769dfe: docs(changelog): resume the changelog and describe 0.34.0 (#2896) (@marmos91)
- c66a3f2: docs(planning): fold the LocalStore deletion and the metrics rethink into Wave 4 (#2858) (@marmos91)
- 95206bf: docs(planning): fold the open issues that bear on the block plan into it (#2831) (@marmos91)
- 824a88b: docs(planning): plan the pkg/block simplification, and record why the checkpoint died (#2828) (@marmos91)
- 5f5024c: refactor(block): Wave 1 — pure deletion across journal, engine, stores and root (#2842) (@marmos91)
- 5750fe6: refactor(block): Wave 2 — extract pkg/block/gc, group the decorators under middleware (#2846) (@marmos91)
- 71ffb2e: refactor(block): Wave 3 — split syncer.go and reclaim.go, withdraw 3A (#2855) (@marmos91)
- 1de0fc3: refactor(block): cut the Store back-reference and the BlockSize alias (#2839) (@marmos91)
- b167091: refactor(config): use maintained mapstructure decode hooks (#2867) (@badrkamel)
- be85d27: refactor(middleware): compose block transforms as a pipeline of stages (#2850) (@marmos91)
- 38fa6d5: test(badger): gate the sparse-lookup cost, and repair the marker it moved (#2892) (@marmos91)
- 52f7cc6: test(bench): cover embedded Samba re-export configuration (#2865) (@badrkamel)
- ee95cc3: test(rpc): fuzz AUTH_UNIX decoding against bounded XDR oracle (#2863) (@badrkamel)
Full Changelog: v0.33.0...v0.34.0
Verifying Checksums
sha256sum -c checksums.txtVerifying Signatures
Release checksums are signed with Sigstore cosign using keyless OIDC:
cosign verify-blob \
--bundle checksums.txt.sigstore.json \
--certificate-oidc-issuer https://token.actions.githubusercontent.com \
--certificate-identity-regexp '^https://github\.com/marmos91/dittofs/\.github/workflows/release\.yml@refs/tags/' \
checksums.txtDocker Images
Multi-architecture images are available for linux/amd64 and linux/arm64:
marmos91c/dittofs:v0.34.0- This specific versionmarmos91c/dittofs:latest- Latest stable releasemarmos91c/dittofs:v0- Latest v0.x.xmarmos91c/dittofs:v0.34- Latest v0.34.x
v0.33.0
DittoFS v0.33.0
A modular virtual filesystem that decouples file interfaces from storage backends.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.33.0/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.33.0/dfs_0.33.0_amd64.pkg.tar.zst
sudo pacman -U dfs_0.33.0_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.33.0Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.33.0/dfs_0.33.0_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.33.0/dfsctl_0.33.0_Linux_x86_64.tar.gz | tar xzChangelog
Bug Fixes
- a9e72a9: fix(block): join in-flight uploads on the carve error path (#2812) (@marmos91)
- 1b61208: fix(block): make the upload window the whole bound on block PUTs (#2780) (@marmos91)
- 8243e9a: fix(block): report uploads in flight, and why an evict reclaimed nothing (#2821) (@marmos91)
- 28468eb: fix(dfsctl): ask for each S3 field the operator did not supply (#2819) (@marmos91)
- d17936b: fix(dfsctl): keep the S3 target echo out of machine output (#2827) (@marmos91)
- 1988d5e: fix(gc): refuse to reap stranded rows from an incomplete live set (#2798) (@marmos91)
- 6ee81c6: fix(metadata): refund the quota charge on a type change, and let an operator see drift before repairing it (#2825) (@marmos91)
- 59cefce: fix(metadata): reject a create whose attr.Type contradicts the entry point (#2740) (#2794) (@marmos91)
- c0402de: fix(metadata): stop dropping undecodable rows from ListFileChunks (#2800) (@marmos91)
- 9a99616: fix(nfsv3): refuse operations on a disabled share (#2816) (@marmos91)
- ac8838a: fix(nfsv4): re-derive the callback verdict when a binding changes (#2799) (@marmos91)
- 02325ab: fix(nfsv4): walk every back-bound connection on CB_RECALL, and keep the current failure (#2806) (@marmos91)
- 4dcbd89: fix(nix): make the cache's entry point repairable, and verify it (#2811) (@marmos91)
- 4c5a7db: fix(runtime): close the block stores before the metadata stores at shutdown (#2802) (@marmos91)
- 97aa7ad: fix(runtime): reach the trash reaper and block GC on the real shutdown path, and fence share publication behind it (#2824) (@marmos91)
Performance Improvements
- 1da70c9: perf(metadata): persist quota counters so startup never scans every file row (#2809) (@marmos91)
- 65bbda7: perf(wave7): run phase 0 locally, and find the group commit limited by shard fan-out (#2818) (@marmos91)
Others
- 2dfca5e: chore(release): v0.33.0 (@marmos91)
- f7c1ee0: ci(nix): host the binary cache on Cubbit DS3 (#2792) (@marmos91)
- d49bb7a: ci(workflows): keep push runs, and name the failing test in the Windows summary (#2797) (@marmos91)
- 6ea886d: ci: build the tree two green PRs would produce (#2823) (@marmos91)
- f0a8f32: docs(planning): re-measure the three SMB boundary cuts, and correct the entry condition (#2814) (@marmos91)
- 6d36441: docs(planning): strike the two landed auxsvc items from the roadmap (#2801) (@marmos91)
- 01b8362: docs(smb): record why the reconnect ClientGuid gate is lease-scoped (#2795) (@marmos91)
- 4d61139: refactor(adapter): fold the three copies of the runtime test fixture into one (#2803) (@marmos91)
- 5047076: refactor(block): declare local-tier capabilities on LocalStore (#2779) (@marmos91)
- e617f7c: refactor(block): drop ContentHash's base64 read path; keep the three load-bearing tolerances (#2777) (@marmos91)
- c03a345: refactor(smb): move CreateContext and TreeConnection into the types package (#2783) (@marmos91)
- a8276c8: refactor(smb): move OpenFile and OpenName into the types package (#2785) (@marmos91)
- b556ca8: test(api): bind ephemeral ports and prove the server under test answered (#2820) (@marmos91)
- 867588e: test(metadata): drop the share-cache assertion that cannot fail (#2786) (@marmos91)
Full Changelog: v0.32.0...v0.33.0
Verifying Checksums
sha256sum -c checksums.txtVerifying Signatures
Release checksums are signed with Sigstore cosign using keyless OIDC:
cosign verify-blob \
--bundle checksums.txt.sigstore.json \
--certificate-oidc-issuer https://token.actions.githubusercontent.com \
--certificate-identity-regexp '^https://github\.com/marmos91/dittofs/\.github/workflows/release\.yml@refs/tags/' \
checksums.txtDocker Images
Multi-architecture images are available for linux/amd64 and linux/arm64:
marmos91c/dittofs:v0.33.0- This specific versionmarmos91c/dittofs:latest- Latest stable releasemarmos91c/dittofs:v0- Latest v0.x.xmarmos91c/dittofs:v0.33- Latest v0.33.x
v0.32.0
DittoFS v0.32.0
A modular virtual filesystem that decouples file interfaces from storage backends.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.32.0/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.32.0/dfs_0.32.0_amd64.pkg.tar.zst
sudo pacman -U dfs_0.32.0_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.32.0Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.32.0/dfs_0.32.0_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.32.0/dfsctl_0.32.0_Linux_x86_64.tar.gz | tar xzChangelog
Features
- eed3469: feat(docker): ship the compose stack the install guide documents (#2616) (@marmos91)
- 0164002: feat(engine): sink upload semaphore, WithoutCancel reap, shutdown test port (@marmos91)
- 998254d: feat(journal): FlushAll drain, per-file closures, carver cut-cursor fix (@marmos91)
- f239a81: feat(journal): lane F engine closure, ported pins, carve retirement (@marmos91)
Bug Fixes
- 1744228: fix(api): answer 409 for a duplicate metadata store name (#2621) (@marmos91)
- 3f9c359: fix(api): evict the health checker under the name it was cached with (#2635) (@marmos91)
- fc3251e: fix(api): invalidate the adapter auth cache on user and group mutations (#2585) (@marmos91)
- 0d5d4cf: fix(auxsvc): start aux services outside the group lock (#2540) (@marmos91)
- 3ead6d4: fix(block-store): count share references by name as well as UUID, and refuse a taken rename before writing (#2612) (@marmos91)
- 3893a39: fix(controlplane): drain API handlers before store close (#2722) (@marmos91)
- 8996031: fix(controlplane): normalize user-route share keys and reproject the root ACL on an id change (#2723) (@marmos91)
- 42e952d: fix(controlplane): project the share root ACL on the grant write, not at each HTTP call site (#2679) (@marmos91)
- c411e22: fix(dfs): name the release that can still migrate a pre-journal share (#2563) (@marmos91)
- c49f4d7: fix(docs): describe the single block store and journal consistently (#2607) (@marmos91)
- 986d828: fix(e2e): make the MinIO availability guard verify the image is obtainable (#2715) (@marmos91)
- 9485651: fix(engine): copilot panel findings (@marmos91)
- 4fa692c: fix(engine): local-only flush populates manifest, matching carve semantics (@marmos91)
- 01b84fd: fix(engine): the upload window never engaged (#2556) (@marmos91)
- a708809: fix(engine): unhealthy remote is the soft condition, matching carve acks (@marmos91)
- 7849328: fix(journal): verify warm reads unless a caller opts out (#2572) (@marmos91)
- e5a87de: fix(lock): recall an NFSv4 delegation on the SMB lock path instead of denying (#2756) (@marmos91)
- b01d10e: fix(metadata): cross-share LINK/RENAME reports one code across protocols (#2536) (@marmos91)
- c66a38f: fix(metadata): preserve DOS attribute bits across an absolute Mode write (#2752) (@marmos91)
- cd256ba: fix(metadata): run postgres transactions at REPEATABLE READ, and write the row the transaction read (#2707) (@marmos91)
- 7bfb2c2: fix(metadata): write back the attribute columns SetFileAttributes was asked for, not the ones that differ from its pre-read (#2711) (@marmos91)
- 6a9240e: fix(netlogon): retry a transient connect failure so hot-reload is transparent (#2732) (@marmos91)
- 4f66683: fix(nfs): advertise only the auth flavors a share accepts (#2578) (@marmos91)
- 13c6139: fix(nfs): apply a register-with-system flip that lands mid-transition (#2662) (@marmos91)
- 83f0b08: fix(nfs): bound the callback AUTH_SYS credential to its RFC 5531 limits (#2724) (@marmos91)
- 3d7cb7f: fix(nfs): cap RPC and MOUNT decodes at the bytes the peer delivered (#2552) (@marmos91)
- b8abc53: fix(nfs): deliver a Kerberos principal's SID and group SIDs to NFS ACL evaluation (#2747) (@marmos91)
- 122cc04: fix(nfs): deny an RPCSEC_GSS credential that was never verified (#2575) (@marmos91)
- bff40e9: fix(nfs): dereference the sidecar config bools under configMu (#2565) (@marmos91)
- 77682e5: fix(nfs): enforce the export auth-flavor policy on every NFSv3 operation (#2584) (@marmos91)
- 6ae6e59: fix(nfs): fail closed when share-permission resolution errors (#2574) (@marmos91)
- 7069013: fix(nfs): gate both share-handle entry points on the export policy (#2403) (#2681) (@marmos91)
- 4e51ad5: fix(nfs): gate the four state-only ops on the export auth policy (#2721) (@marmos91)
- b2b010b: fix(nfs): grant and recall NFSv4.1 delegations over a shared connection (#2598) (@marmos91)
- 55861c2: fix(nfs): guard sidecar config fields across settings-apply paths (#2549) (@marmos91)
- e9287d6: fix(nfs): keep the GSS context LRU within a principal, and touch it only after the MIC (#2716) (@marmos91)
- 9604b5d: fix(nfs): krb5p EC integrity, GSS ticket-bound context lifetime, UDP shutdown wait (#2405) (#2686) (@marmos91)
- dc988ae: fix(nfs): notify dir delegations on create, remove, rename, link (#2537) (@marmos91)
- 4064a4b: fix(nfs): push export policy changes into the running share (#2582) (@marmos91)
- e2f09cc: fix(nfs): recall an NFSv4 delegation before foreign I/O, and complete the reclaim (#2746) (@marmos91)
- fbe525c: fix(nfs): refuse a persisted require_kerberos when Kerberos is decommissioned (#2594) (@marmos91)
- 531e002: fix(nfs): refuse an export that would accept no auth flavor (#2639) (@marmos91)
- 4e603ab: fix(nfs): release callback state by binding direction, and walk past two candidates (#2727) (@marmos91)
- bbbeaf4: fix(nfs): release callback waiters by session, not only by connection (#2650) (#2742) (@marmos91)
- ba5a388: fix(nfs): return NFS4ERR_DELAY when a v4.1 session is at its connection limit (#2590) (@marmos91)
- a105464: fix(nfs): scope UMNT to one export and UMNTALL to the calling client (#2583) (@marmos91)
- d8ac11f: fix(nfs): seek answers the file type, not a blanket ISDIR (#2538) (@marmos91)
- 33b681f: fix(nfs): serialise system-rpcbind registration against shutdown (#2696) (#2718) (@marmos91)
- d9a1032: fix(nfs): start the v4.1 recall watchdog at dequeue, bound the queue wait separately (#2728) (@marmos91)
- 03fbb51: fix(nfs): sync sidecar shutdown state outside the group lock (#2547) (@marmos91)
- 9486794: fix(nfs): validate the v4.0 lock-owner seqid of zero (#2541) (@marmos91)
- 04da056: fix(nfsv3): SETATTR block reclaim on partial failure, stale-handle status, write verifier resolution (#2683) (@marmos91)
- 24eb879: fix(nfsv4): index the v4.1 owner map with the conversion inline (#2596) (@marmos91)
- 06a6f26: fix(nfsv4): write the client-recovery row when a client takes state, and retire it at grace end (#2689) (@marmos91)
- 8da6b69: fix(nlm): authorize byte-range locks, dedup retransmits, and own the NLM/NSM background work (#2690) (@marmos91)
- ...
v0.31.1
DittoFS v0.31.1
A modular virtual filesystem that decouples file interfaces from storage backends.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.31.1/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.31.1/dfs_0.31.1_amd64.pkg.tar.zst
sudo pacman -U dfs_0.31.1_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.31.1Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.31.1/dfs_0.31.1_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.31.1/dfsctl_0.31.1_Linux_x86_64.tar.gz | tar xzChangelog
Features
- 02f2be1: feat(nfs): grant delegations to NFSv4.1 clients on a verified callback path (#2479) (@marmos91)
Bug Fixes
- ecdb4ed: fix(api): drop RealIP so the request log records an unforgeable address (#2291) (@marmos91)
- f793f27: fix(block): hold the retention pin in the local store so Start cannot lift it (#2302) (@marmos91)
- 4ac5071: fix(block): never evict bytes a share has no way to fetch back (#2376) (@marmos91)
- 29ffe1d: fix(block): route the carve block size to the loop that reads it (#2311) (@marmos91)
- 9314a04: fix(block): stop reporting a local-tier check that never ran (#2347) (@marmos91)
- 2406afb: fix(ci): analyse cold so a bumped go.sum cannot poison the lint cache (#2443) (@marmos91)
- 8b05dc1: fix(conformance): stop reporting a setup failure as new test failures (#2477) (@marmos91)
- 5454e4f: fix(deps): bump grpc to 1.83.1 in the operator module (#2411) (@marmos91)
- a2791f9: fix(journal): fail a restore whose shard can never be made durable (#2390) (@marmos91)
- d3172e6: fix(journal): fence a hydrate against the truncate marker's own version (#2386) (@marmos91)
- ad6162b: fix(journal): order unstamped segments by age, not ahead of everything (#2308) (@marmos91)
- 90a7b8d: fix(journal): record which writer made a cold entry, and stop restore guessing (#2478) (@marmos91)
- 58dc7c2: fix(journal): size the carve block arena to the share's own chunk ceiling (#2480) (@marmos91)
- bc58034: fix(lint): stop the citation check scanning nested working copies (#2267) (@marmos91)
- a02d55f: fix(metadata): make the xattr read-modify-write atomic (#2452) (@marmos91)
- 111192f: fix(metadata): persist the create verifier on SQL backends, and rename the row the transaction read (#2379) (@marmos91)
- 9f35691: fix(metadata): release a file's bytes from the share usage counters when its last link goes (#2288) (@marmos91)
- cde7dd9: fix(metadata): rename over an existing file never freed the clobbered file's bytes (#2309) (@marmos91)
- 2a3b305: fix(nfs): DESTROY_CLIENTID must not gate on the requester's client ID, and must report an unrecognized target first (#2489) (@marmos91)
- 1d69f4d: fix(nfs): bind NFSv4 stateids to their owning client and make the stateid other unpredictable (#2448) (@marmos91)
- 48671a2: fix(nfs): call the cross-protocol lock manager without holding the global state mutex (#2459) (@marmos91)
- 47a439e: fix(nfs): check read permission in NFSv4 READ, READ_PLUS and SEEK (#2427) (@marmos91)
- 5f05f50: fix(nfs): check the stateid belongs to the caller in CLOSE, LOCK, LOCKU, OPEN_DOWNGRADE, OPEN_CONFIRM and TEST_STATEID (#2466) (@marmos91)
- 1a478fc: fix(nfs): default max_version to 4.0 until full NFSv4.1 lands (@marmos91)
- 06cef31: fix(nfs): draw the server boot epoch at random instead of reading the clock (#2485) (@marmos91)
- c5dd09b: fix(nfs): enforce CREATE_SESSION channel-size floors, flag validation, and NOSPC session cap (#2471) (#2494) (@marmos91)
- 13931d0: fix(nfs): implement the EXCHANGE_ID flag validation and record algorithm of RFC 8881 18.35.4 (#2491) (@marmos91)
- f9ebda7: fix(nfs): key the RECLAIM_COMPLETE duplicate check on the client, not the grace period (#2486) (@marmos91)
- c810d03: fix(nfs): merge same-owner byte-range locks and keep one lease per client ID (#2361) (@marmos91)
- 1cee59f: fix(nfs): narrow the SECINFO flavor list to the target share's auth-flavor policy (#2460) (@marmos91)
- 2acb10d: fix(nfs): reject bad, old and stale stateids in CLOSE, OPEN_CONFIRM, OPEN_DOWNGRADE and LOCK (#2484) (@marmos91)
- a4b0ee9: fix(nfs): reject cross-share LINK — the v3 share check was a tautology (#2432) (@marmos91)
- 14fd86a: fix(nfs): return NFS4ERR_NOT_ONLY_OP when a session-exempt op shares its COMPOUND (#2481) (@marmos91)
- 321a443: fix(nfs3): narrow ACCESS replies to the request and drop DELETE for non-directories (#2273) (@marmos91)
- a8c8772: fix(nfs4): translate the all-ones lock length into the manager's end-of-file range (#2274) (@marmos91)
- 655f12b: fix(nfsv4): REMOVE never freed the removed file's payload bytes (#2286) (@marmos91)
- fc16cb0: fix(nfsv4): put space_avail/free/total at their RFC 7530 attribute numbers (#2458) (@marmos91)
- 2a6ac1b: fix(postgres): create the filesystem_meta table the share queries read (#2316) (@marmos91)
- b608506: fix(smb): answer a CHANGE_NOTIFY that registers after a directory is marked deleted (#2283) (@marmos91)
- 93c8116: fix(smb): authorize the parent-dir frozen-timestamp restore by the frozen handle's grant (#2319) (@marmos91)
- 99d2203: fix(smb): bind a client-supplied FileId to the session that opened it (#2461) (@marmos91)
- 24846cb: fix(smb): correct the seven wrong spec citations and empty the allowlist (#2269) (@marmos91)
- ee3ab6a: fix(smb): don't announce a rename that a hard link made a no-op (#2387) (@marmos91)
- 923c40e: fix(smb): hardlink replace never freed the replaced file's bytes (#2318) (@marmos91)
- 63b1aa4: fix(smb): keep the ChangeTime a client observed across rename (#2284) (@marmos91)
- c77afc8: fix(smb): let FILE_WRITE_ATTRIBUTES authorize a timestamp write (#2313) (@marmos91)
- a39ec44: fix(smb): let rename's ChangeTime advance stand, and pin frozen timestamps across rename + CLOSE (#2279) (@marmos91)
- c5ba24c: fix(smb): make the rename ChangeTime preserve conditional (#2343) (@marmos91)
- 0bc01b3: fix(smb): scope async-flagged CANCEL lookups to the arriving connection (#2426) (@marmos91)
- 9295779: fix(smb): scope the tree lookup to the requesting session (#2424) (@marmos91)
- 3c22a38: fix(smbtorture): anchor the NT_STATUS_NO_MEMORY flake excuse to a connection diagnostic (#2272) (@marmos91)
- 025de21: fix(test): stop the known-failures parser reading fenced examples as rows (#2412) (@marmos91)
Performance Improvements
- a9a429e: perf(journal): read the file high-water mark off the interval tail (#2384) (@marmos91)
- 8212870: perf(metadata): let a directory listing skip the attributes nobody reads (#2292) (@marmos91)
Others
v0.31.0
DittoFS v0.31.0
A modular virtual filesystem that decouples file interfaces from storage backends.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.31.0/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.31.0/dfs_0.31.0_amd64.pkg.tar.zst
sudo pacman -U dfs_0.31.0_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.31.0Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.31.0/dfs_0.31.0_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.31.0/dfsctl_0.31.0_Linux_x86_64.tar.gz | tar xzChangelog
Features
- 339a134: feat(integrity): run the structural manifest scan on a schedule (#2080) (@marmos91)
- dd82b32: feat(lint): fail CI when a comment names a structure and cites a different spec section (#2192) (@marmos91)
- ca01f58: feat(offline): report per-share remote-only bytes and offline read safety (#2082) (@marmos91)
Bug Fixes
- 44f6aaa: fix(adapters): an out-of-range port panicked under the service lock and wedged the control plane (#2213) (@marmos91)
- 61dcaa3: fix(api): gate the drain-uploads stall trip on there being something to upload (#2099) (@marmos91)
- e87b695: fix(badger): apply capabilities in memory only after the write commits (#2221) (@marmos91)
- 54acb39: fix(block): bound parsed chunk offsets to int64 at the parse (#2118) (@marmos91)
- 4a2c430: fix(block): give a manifest row an in-chunk start offset so it can be narrowed off its head (#2137) (@marmos91)
- 3ef4d7d: fix(block): order the carve dedup oracle against the remote GC sweep (#2249) (@marmos91)
- 2e40487: fix(block): seed the local tier for a server-side copy's destination (#2161) (@marmos91)
- bf03c41: fix(block): stop a reflink's destination serving the content it replaced (#2178) (@marmos91)
- 7f1b811: fix(carve): keep a row reaching past a reaped run, batch the reap, evict under carveMu (#2122) (@marmos91)
- 0760d97: fix(carve): keep what a run's first chunk takes over, where the journal says it is owed (#2163) (@marmos91)
- 0d3aab2: fix(clone): clip a local-only clone's destination past the source's size (#2188) (@marmos91)
- a675d03: fix(dfsbench): judge the cold-barrier drain residue by the segments it pins (#2088) (@marmos91)
- 71130f8: fix(engine): cross-check offline readiness against the manifest (#2121) (@marmos91)
- f6038a6: fix(engine): hold the per-file commit lock while reaping superseded manifest rows (#2071) (@marmos91)
- d9847e6: fix(journal): fence Hydrate against a racing Delete (#2252) (@marmos91)
- 4a56382: fix(journal): fold the cold log into the version RestoreToVersion rewinds to (#2253) (@marmos91)
- 4906e43: fix(journal): fsync the view RestoreToVersion re-materializes (#2248) (@marmos91)
- 85e1314: fix(journal): let the write-path capacity gate force-seal synced actives (#2266) (@marmos91)
- 72f0b25: fix(journal): persist Invalidate's cold demotion before marking the interval (#2103) (@marmos91)
- 82e7e18: fix(journal): repack target never counted its records, so eviction could drop dirty bytes (#2246) (@marmos91)
- 555fdfc: fix(journal): truncate the torn cold.log tail at recovery (#2245) (@marmos91)
- db68d21: fix(metadata): make DeleteChild idempotent across all four backends (#2261) (@marmos91)
- 0692f19: fix(nfs4): bind the client ID to the authenticating principal where the RFCs require it (#2173) (@marmos91)
- dd3477b: fix(nfs4): deliver delegation recalls — AUTH_SYS credential and the client's callback_ident (#2219) (@marmos91)
- 69a5689: fix(nfs4): key v4.1 open- and lock-owners by the session's client (#2146) (@chessman)
- ff14b24: fix(nfs4): open the reboot-grace window only when state is reclaimable (#2214) (@marmos91)
- 60e2bd9: fix(nfs4): reject the byte ranges RFC 7530 calls invalid in LOCK (#2210) (@chessman)
- f79801f: fix(nfs4): report only the ACCESS rights the object and minorversion can answer (#2211) (@chessman)
- 8f614a9: fix(nfs4): validate the client ID on OPEN (#2063) (@chessman)
- 04be48b: fix(shares): reject share names that cannot address their own directory (#2116) (@marmos91)
- 01aadfe: fix(smb): answer a CHANGE_NOTIFY from events that already exist (#2138) (@marmos91)
- 23b9742: fix(smb): answer every CHANGE_NOTIFY the notify path owes, one operation at a time (#2175) (@marmos91)
- 05062d1: fix(smb): answer every CHANGE_NOTIFY the registry removes (#2131) (@marmos91)
- b031b1c: fix(smb): apply the three SET_INFO access gates MS-FSA states as MUSTs (#2184) (@marmos91)
- e3da5b9: fix(smb): arm the session-cleanup barrier before draining a dying connection (#2120) (@marmos91)
- 73f3ef3: fix(smb): break a deleted file's Handle leases only once teardown actually removed it (#2183) (@marmos91)
- 47a2d1f: fix(smb): decide content-change dir-lease breaks at change time, not at dispatch time (#2081) (@marmos91)
- 1dedf7d: fix(smb): decide the last delete-on-close handle atomically with its departure (#2143) (@marmos91)
- 4239639: fix(smb): decline a delete-on-close removal of a non-empty directory instead of failing the close (#2159) (@marmos91)
- 8850852: fix(smb): dispatch FSCTL_SET_REPARSE_POINT on the code clients actually send (#2194) (@marmos91)
- cf184ba: fix(smb): gate rename on the destination parent's write sharing, not delete sharing (#2135) (@marmos91)
- 5dda058: fix(smb): give buffered notify events to the earliest outstanding request (#2151) (@marmos91)
- 57fb21f: fix(smb): key per-lease state on the lease's identity, not on the lease key alone (#2139) (@marmos91)
- 865b92e: fix(smb): lease epoch only seeds on first grant; rename break precedes the destination lookup (#2124) (@marmos91)
- d1ef833: fix(smb): order response emission per connection so a break is never overtaken (#2162) (@marmos91)
- 6141384: fix(smb): relate base files and streams by resolved identity, not path text (#2141) (@marmos91)
- feb2ea5: fix(smb): release the DH2Q replay reservation before the parked CREATE's final response (#2097) (@marmos91)
- a013628: fix(smb): scope a lease's epoch and state to its own file, not the lease key (#2149) (@marmos91)
- 278b133: fix(smb): scope lease ack, reclaim and break staging to the lease's own file (#2157) (@marmos91)
- 9cdaeeb: fix(smb): stop bumping LastWriteTime on an attribute-only SET_INFO (#2191) (@marmos91)
Performance Improvements
- 4f2fe25: perf(journal): carve a file's dirty runs concurrently (#2072) (@marmos91)
- 5d1f6d8: perf(journal): pack one remote block across several dirty runs (#2092) (@marmos91)
- b132ff1: perf(seed): batch the manifest cold seed into one durable append (#2087) (@marmos91)
- 6d2afd6: perf(shares): read only the size the share-start reconcile compares (#2077) (@marmos91)
- 1c7b735: perf(sql): seek file_blocks by ID range instead of scanni...
v0.30.0
DittoFS v0.30.0
A modular virtual filesystem that decouples file interfaces from storage backends.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.30.0/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.30.0/dfs_0.30.0_amd64.pkg.tar.zst
sudo pacman -U dfs_0.30.0_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.30.0Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.30.0/dfs_0.30.0_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.30.0/dfsctl_0.30.0_Linux_x86_64.tar.gz | tar xzChangelog
Features
- 632bbf0: feat(dfsctl): add store check, a metadata-only manifest coverage scan (#2008) (@marmos91)
- 5204481: feat(dfsctl): repair the recoverable manifest damage store check reports (#2044) (@marmos91)
- 40ee8d3: feat(encryption): support master-key rotation via a retired-key set (#1951) (@marmos91)
- e732094: feat(kmip): validate the client against PyKMIP in CI and honour the KMIP key state (#2009) (@marmos91)
- 2f4b71c: feat: report why a share is not being served (#2011) (@marmos91)
Bug Fixes
- ae309f9: fix(adapters): report the real bind outcome from adapter enable (#2045) (@marmos91)
- 766c6ed: fix(api): bind {scope} on the id-less quota routes (#2055) (@marmos91)
- 5a2e620: fix(audit): close the LOW bug and security findings across block, metadata and control plane (#1971) (@marmos91)
- e411551: fix(authz): close three data-plane authorization gaps (#1932) (@marmos91)
- f070b6a: fix(badger): resolve a missing link-count key to the type-based default (#1989) (@marmos91)
- d14a30b: fix(block): block the legacy cas/ purge when a configured share is not registered (#1955) (@marmos91)
- 03779e3: fix(block): fetch chunks after a sparse hole instead of skipping to the block boundary (#1903) (@marmos91)
- f7f445b: fix(block): hold a cold-read hydrate to the bytes its manifest row still claims (#1933) (@marmos91)
- 16ff798: fix(block): journal reclaim/recovery correctness and local-store cleanups (#1948) (@marmos91)
- a912262: fix(block): make the covering-row check overflow-safe (#1969) (@marmos91)
- 3d65945: fix(block): re-seed cold intervals a share lost under an earlier build (#1889) (@marmos91)
- 7d66d61: fix(block): reclaim the legacy cas/ namespace per remote, not per share (#1928) (@marmos91)
- 949c0c3: fix(block): reconcile an uncovered read against the manifest, not just a cold one (#1965) (@marmos91)
- 1f95b38: fix(block): report the physical segment footprint as local disk_used (#1890) (@marmos91)
- 8c29143: fix(block): resolve overlapping manifest rows to the greatest start (#1953) (@marmos91)
- 8c41f31: fix(block): share decorator passthrough, drop dead append contract, correct rotation docs (#1946) (@marmos91)
- a4e95ab: fix(block): stage whole blocks on readahead, fail closed on a cold re-read (#1959) (@marmos91)
- 67a5890: fix(block): stop a stale straddling manifest row shadowing a newer row on cold reads (#1991) (@marmos91)
- 7f221d5: fix(block): verify record CRCs before copying payloads to remote or a new segment (#1925) (@marmos91)
- 336743c: fix(block): write manifest rows for deduped carve chunks (#1979) (@marmos91)
- c745cbe: fix(controlplane): validate the control-plane config before the first-boot admin bootstrap (#2056) (@marmos91)
- e72e068: fix(gc): propagate remote-tier error detail into the server-wide sweep total (#2040) (@marmos91)
- e88d008: fix(journal): a hydrate fills the local tier, it does not overwrite it (#2064) (@marmos91)
- b5af5ce: fix(journal): bound the age of unfsynced writes with a periodic dirty-age commit (#2005) (@marmos91)
- ace69fe: fix(journal): carve through the manifest row a run's end lands inside (#2049) (@marmos91)
- 083f43f: fix(journal): cover the framed FileID with the record CRC (#1988) (@marmos91)
- a3db32b: fix(journal): release the segment claim when eviction's cold-log append fails (#1922) (@marmos91)
- 2f5205e: fix(journal): skip a sealed segment that scans as zero valid records instead of panicking (#1921) (@marmos91)
- 99767ae: fix(lifecycle): abort startup when the machine SID is not durable (#1987) (@marmos91)
- adb8bf0: fix(lock): close the wait-for-graph check-then-act window and bound two lease-key sweeps (#1935) (@marmos91)
- 0131bac: fix(lock): derive a lease record's lock type, and take the store round-trip out from under the share mutex (#2043) (@marmos91)
- f175047: fix(memory): guard durable handles with the store-wide lock (#2034) (@marmos91)
- 87e9587: fix(metadata): bound the backpressure guard by the store's retry budget, not wall clock (#2068) (@marmos91)
- af7086f: fix(metadata): close five MED write-path audit findings (#1972) (@marmos91)
- ace4a92: fix(metadata): close small MED audit findings in the memory and badger stores (#1970) (@marmos91)
- 2d37dc7: fix(metadata): close the MED audit findings in the SQL stores (#1949) (@marmos91)
- a32bec1: fix(metadata): copy every FileAttr field in CopyFileAttr (#1982) (@marmos91)
- 46a96c9: fix(metadata): drop every derived badger cache on Reset and RestoreSnapshot (#2004) (@marmos91)
- f3634ae: fix(metadata): enforce per-share quotas against the share's own usage, not the metadata store's total (#2066) (@marmos91)
- 537f390: fix(metadata): make Identity.HasGID a pure read so concurrent callers are safe (#1902) (@marmos91)
- fadd83f: fix(metadata): make ListFileChunks membership agree across backends (#2046) (@marmos91)
- f560213: fix(metadata): narrow the straddling manifest row a carve run supersedes (#2039) (@marmos91)
- fcf560a: fix(metadata): purge the badger share cache on Reset and RestoreSnapshot (#1950) (@marmos91)
- 7cd665b: fix(metadata): reject an undecodable handle in statfs on every backend (#2069) (@marmos91)
- d465318: fix(metadata): scope durable-handle FileID lookups to a single handle (#1926) (@marmos91)
- e582dd8: fix(nfs): correct non-regular READ errno, clamp v3 READ to rtmax (#1939) (@marmos91)
- 78ddd44: fix(nfs): deny NFSv4 delegations over cross-protocol byte-range locks (#1990) (@marmos91)
- f93b7d2: fix(nfs): enforce netgroup export access on NFSv4, not only NFSv3 MOUNT (#1907) (@marmos91)
- 0a253f8: fix(nfs): register with the system rpcbind without an adapter restart (#1966) (@marmos91)
- 7679499: fix(nfs): report the real NFS version for a connected client (#2007) (@marmos91)
- 29928f1: fix(nfs4): consume the owner seqid when a sequenced operation fails (#1942) (@chessman)
- 5901017: fix(nfs4): release open state when a v4.1 client's lease expires (@chessman)
- 980c302: fix(runtime): surface adapter restart failures, hold entries until stopped, drop dead client tracking (#1943) (@marmos91)
- 7c93ca2: fix(s3): enforce the endpoint sa...
v0.29.2
DittoFS v0.29.2
A modular virtual filesystem that decouples file interfaces from storage backends.
Installation
Quick install (macOS / Linux):
curl -fsSL https://github.com/marmos91/dittofs/releases/download/v0.29.2/install.sh | shHomebrew (macOS/Linux):
brew tap marmos91/tap
brew install marmos91/tap/dfs # Server daemon
brew install marmos91/tap/dfsctl # Client CLIDebian/Ubuntu (APT):
curl -fsSL https://s3.cubbit.eu/dittofs-binaries/apt/dittofs.gpg.key | gpg --dearmor --yes | sudo tee /usr/share/keyrings/dittofs.gpg > /dev/null
echo "deb [signed-by=/usr/share/keyrings/dittofs.gpg] https://s3.cubbit.eu/dittofs-binaries/apt stable main" | sudo tee /etc/apt/sources.list.d/dfs.list
sudo apt update && sudo apt install dfs
sudo systemctl enable --now dfsRHEL/Fedora (YUM):
sudo curl -fsSLo /etc/yum.repos.d/dfs.repo https://s3.cubbit.eu/dittofs-binaries/rpm/dfs.repo
sudo yum install dfs
sudo systemctl enable --now dfsArch Linux:
curl -fsSLO https://github.com/marmos91/dittofs/releases/download/v0.29.2/dfs_0.29.2_amd64.pkg.tar.zst
sudo pacman -U dfs_0.29.2_amd64.pkg.tar.zst
sudo systemctl enable --now dfsScoop (Windows):
scoop bucket add marmos91 https://github.com/marmos91/scoop-bucket
scoop install dfs # Server daemon
scoop install dfsctl # Client CLIDocker:
docker run -d \
-p 12049:12049 \
-p 12445:12445 \
-p 8080:8080 \
-v /path/to/config.yaml:/config/config.yaml:ro \
-v dittofs-data:/data \
marmos91c/dittofs:v0.29.2Binary Download:
# dfs (server daemon)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.29.2/dfs_0.29.2_Linux_x86_64.tar.gz | tar xz
# dfsctl (client CLI)
curl -sfL https://github.com/marmos91/dittofs/releases/download/v0.29.2/dfsctl_0.29.2_Linux_x86_64.tar.gz | tar xzChangelog
Bug Fixes
- d41cf8f: fix(badger): index legacy rows by payload once, instead of scanning per lookup (#1886) (@marmos91)
- efb864c: fix(block): refuse a manifest row that cannot be placed, instead of reading it as a hole (#1881) (@marmos91)
- e9c639a: fix(block): report upload progress per block, not per carve pass (#1874) (@marmos91)
- b5c01e6: fix(postgres): guard the manual migrate path against a future schema (#1870) (@marmos91)
- 4f24067: fix(smb): return FiletimeToTime in UTC, not local time (#1884) (@viquitorreis)
- 67c9a0a: fix(stores): stamp the on-disk format and refuse state from a newer release (#1865) (@marmos91)
Others
- cb11243: Report what the drain observed, and assert a carve leaves nothing unsynced (#1878) (@marmos91)
- da7a0b5: chore(release): v0.29.2 (@marmos91)
- 55d28e5: docs(bench): publish the cold-read matrix (#1876) (@marmos91)
- 9836716: test(bench): report unsynced size while a drain is in flight (#1880) (@marmos91)
- 7c48a71: test(bench): say when the cold barrier could not verify itself (#1871) (@marmos91)
- 39b39d7: test(cli): scope the drift scan to the visible tree, and prove it scanned it (#1863) (@marmos91)
Full Changelog: v0.29.1...v0.29.2
Verifying Checksums
sha256sum -c checksums.txtVerifying Signatures
Release checksums are signed with Sigstore cosign using keyless OIDC:
cosign verify-blob \
--bundle checksums.txt.sigstore.json \
--certificate-oidc-issuer https://token.actions.githubusercontent.com \
--certificate-identity-regexp '^https://github\.com/marmos91/dittofs/\.github/workflows/release\.yml@refs/tags/' \
checksums.txtDocker Images
Multi-architecture images are available for linux/amd64 and linux/arm64:
marmos91c/dittofs:v0.29.2- This specific versionmarmos91c/dittofs:latest- Latest stable releasemarmos91c/dittofs:v0- Latest v0.x.xmarmos91c/dittofs:v0.29- Latest v0.29.x