AIStor Server Release Notes

This page documents release notes from the last 6 calendar months of AIStor Server releases.

• make sure to use uncached DiskInfo() to skip drives by @harshavardhana
• combine cleanup of part.N and part.N.meta into RenamePart() by @harshavardhana
• Check error when creating Elastic index by @dvaldivia
• CompleteMultipart scenario with delays across nodes by @shtripat
• grid: Avoid goroutine leak when the handler crashes by @vadmeste
• heal: HealObject to use write lock only during data commit by @vadmeste
• fix: panic moveToTrash when drive is almost full by @jiuker
• Fix wait time regression with batch catalog tests by @shtripat

• Allow users to opt out of the cleanup during decommission by @harshavardhana
• Include license id in cluster registration info by @anjalshireesh
• Implement quicker detection when the bucket is empty, rather than listing by @harshavardhana
• add retries to site replication tests by @poornas
• fix: we should turn-off auditLogging for internal ops in Decom by @harshavardhana
• Add checksum decoder by @klauspost
• Increase wait time for site replication scenarios by @shtripat
• fix bug in encrypted metadata causing JSON unmarshaling to fail by @aead
• Add rotating trash prefixes by @klauspost
• Revert “add faster detection if bucket is empty, instead of listing” + tests by @vadmeste
• tests: increase all integration tests to timeout at 5minutes by @harshavardhana
• fix bug in encrypted metadata causing JSON unmarshaling to fail by @aead

• use non-blocking locks when it makes sense by @harshavardhana
• Record successful reads and writes separately for drive health check by @marktheunissen
• Security fix: Use audience sts.min.io by default to invoke Kubernes API TokenReview. by @ramondeklein
• Move formatting related types and methods into a pkg by @krisis
• ilm: ignore object not found errors during expiration by @krisis
• Allow configuring of grid connPingInterval, to take nodes offline faster than the default by @marktheunissen
• Add back the missing v4/object-summary admin API by @marktheunissen
• s3: DeleteObject to succeed when no enough meta quorum AND one offline node by @vadmeste
• Migrate Elasticsearch client from v7 to v9 by @sivanantham-gnanavel
• deps: Update miniohq/ec to fix a server crash by @vadmeste
• use mpCache support by default and when not S3ExpressZone by @harshavardhana
• avoid serializing callers with iam.CheckKey() write locks by @harshavardhana
• Verify LDAP configurations set for the cluster by @shtripat
• admin: Clean up and revert two accidently removed APIs by @vadmeste
• fallback on quotaConfig from older releases on existing buckets by @harshavardhana
• s3: Add whitspaces to complete multipart upload API response by @vadmeste
• More catalog batch job tests with filters by @shtripat
• replication: add in progress metric and target info per node by @poornas

• Reduce write allocations significantly in certain setups by @klauspost

• Add streaming readers/writers by @klauspost
• Direct msgp to decode time.Time in UTC for all time values by @marktheunissen
• Add Trino User-Agent for spark delete optimization by @poornas
• Skip disabled target(s) for replication config updates by @poornas
• fix checksum validation in AppendObject by @poornas
• When a call to decryptPartsChecksums fails, log the object key that t… by @marktheunissen
• Move storage.DeleteBulk to grid by @klauspost
• Handle big folders better by @klauspost
• Update console package by @cesnietor
• iam: Do not reload at startup unless IAM data are mutated by @vadmeste
• Fix O_DIRECT calls failing due to random buffer alignment & discarded buffers by @klauspost
• scanner: Avoid racy double scanning of the same drive by @vadmeste
• healthcheck: Make sure at least one peer is accessible by @vadmeste
• Add tiered fanout calls by @klauspost
• allow single drive per set for experimentation by @harshavardhana
• make deletes immediate to share the latency with the API caller by @harshavardhana
• feat: support object lock on existing buckets by @jiuker
• Check start address of aligned buffers by @klauspost
• Use license auth in callhome by @anjalshireesh
• update all deps by @harshavardhana
• Fix fanout “unknown mux handler” errors by @klauspost
• repl: Set 15 minutes as a header timeout for replication requests by @vadmeste
• provide non-blocking locks for local setups by @harshavardhana
• add KMS key versions support for KES by @aead
• allow max fd count to arbiter in-case requests per node is too high by @harshavardhana
• Send only registration info in public trial by @anjalshireesh

• Use references to aistor console by @allanrogerr

• heal: Add WriteDists to XL meta version by @vadmeste
• Return a user-friendly error when license in read-only phase by @dhananjaykrutika
• fix delay between time-info of local vs peers by @anjalshireesh
• use Walk() unsorted for DeleteBucket() check by @harshavardhana
• Remove test log by @dhananjaykrutika
• prometheus: Do not count dangling objects removal as healing error by @vadmeste
• Multi pool support by @shtripat
• Fix swapped input/output blocked telemetry by @klauspost
• Use latest automation-go changes by @shtripat
• Fix WriteAlignedTo write side hang on limit=-1 & cleanup by @klauspost
• Do globalNotificationSys.DeleteUploadID async by @klauspost
• Use ec instead of mc by @reivaj05
• apply dynamic api config changes on chg by @poornas
• Correct spelling by @shtripat
• fix: holding locks for trackMultipleObjectVersions by @jiuker
• move from unmaintained jsoniter -> go-json by @harshavardhana
• updating object summary API by @zveinn
• scanner: Speed up when a cluster has hundreds of empty buckets by @vadmeste
• support versions list stack replicate between non-minio S3 and minio by @jiuker
• Introduce new Admin API v4 by @harshavardhana
• Remove timestamp from xl writes distribution by @vadmeste
• Update minio/pkg to latest (v3.1.7) by @anjalshireesh
• config: Fix unquoting configuration value by @vadmeste
• fix hadoop User Agent for spark optimization by @poornas
• add missing telemetry on few replication calls by @harshavardhana
• Add more size interval to ttfb histogram by @poornas
• test-fix: check MinIO health before running mc alias set by @sivanantham-gnanavel
• fix: limit the version enable on both source and target for batch replication by @jiuker
• use common buffer for pathJoin scanner for entire readDir() by @harshavardhana
• fix fmt-gen to default console-address by @poornas
• Add a test for hot drive swapping by @vadmeste
• Allow selective buckets to skip scanning by @poornas
• Add CreateSession API support for S3 Express by @harshavardhana
• Added logic to capture replication diag info by @dhananjaykrutika
• re-do pathJoin to use string builder instead of bufferpool by @harshavardhana

• add S3 Express based unsorted listing by @harshavardhana

• add CommitXL API for safer PUTs on unversioned buckets by @harshavardhana
• add support for SSE Append() by @harshavardhana
• use unsorted walk() for IAM reload, replication resync by @harshavardhana
• Move generation tools to go.mod by @klauspost
• CI: sep. jobs for make test & make test-race by @donatello
• ObjectSummary API Update by @zveinn
• Move plugin authZ and authN to IAMSys by @donatello
• fix tracked delete cleanup by @poornas
• support autogenerated credentials for KMS_SECRET_KEY properly by @harshavardhana
• Fix incorrect skipping of tests in SFTP tests by @donatello
• Add object speedtest tracing by @klauspost
• chore: use incremental port for console instead of random by @sivanantham-gnanavel
• Adding missing configuration help values by @zveinn
• Use more “official” OpenTelemetry keys. by @klauspost
• feat: Add JSON output format to catalog jobs by @donatello
• internodes: Remove healthcheck and use grid to mark nodes offline by @vadmeste
• xl: Re-initialize failed local drives by @vadmeste
• fix: append() on plain-text objects by @harshavardhana
• Enable the format.bin check with latest MinIO by @shtripat
• batch: Update current object when storing in the drives by @vadmeste
• Use latest version of miniohq/license by @anjalshireesh
• cleanup iam: simplify store api + fixes by @donatello
• Remove replace by @klauspost
• Remove unused parameter ‘spent’ by @marktheunissen
• Store cached token revokes in xsync.MapOf by @taran-p
• Add Compression: on|off option for batch catalog by @donatello
• kms: add support for computing MACs with MinKMS by @aead
• Add encrypted buffer reuse by @klauspost
• heal: Avoid disabling scanner healing in single and dist erasure mode by @vadmeste
• actually run integration tests on mint runners by @harshavardhana
• batch: Do not resume a failed job by @vadmeste
• update specific dependencies to fix more relevant bugs by @harshavardhana
• fix: batch expiry job doesn’t report delete marker in batch-status m… by @jiuker
• move away from GitHub CI/CD for extensive tests by @harshavardhana
• s3: Strict matching for null versions during GET and Copy Object by @vadmeste
• Add a postgres timeout of 30s, to avoid hanging and blocking MinIO by @marktheunissen
• fix: return an appropriate error for AppendObject() under regular S3 by @harshavardhana
• Migrate SSEC site replication test to new framework by @shtripat
• sr: fix possible race b/w site removal and bucket replication config … by @poornas
• skip partInfo for all versions in fileInfo serialization by @harshavardhana
• fix: missing validation in IAM load secret key by @donatello
• Update pkg version by @reivaj05
• move some tests out of mint runners by @harshavardhana
• Fix typo that causes update via HTTP to fail by @ramondeklein
• Added a README for integration tests by @shtripat
• Make server stop process a cleanup routine by @taran-p
• Fix openID authentication issue by @taran-p
• Refactor to reuse common code in tests by @shtripat
• sync: allow concurrent aborts on active uploadParts() by @jiuker
• cleanup: Simplify global Erasure/Dist/SD usage by @vadmeste
• heal: Resume healing when there is an error with listing by @vadmeste
• tests: wait for site replication to init after restart by @poornas
• replication: replay last queued entries on node restart by @poornas
• Migrate SSE KMS based site replication test by @shtripat
• Update AIStor Assets to v0.0.10 by @dvaldivia
• fix: networking for mint tests and re-enable them by @harshavardhana
• Test multi site replication with MinIO IDP by @shtripat
• s3: Consistent returning of tagging/object locking information by @vadmeste
• Allow specifying a psql schema with the table name by @ramondeklein

• Bump up to 1.24.2 for security issue by @donatello

• move IAM data loading to the background by @donatello
• implement Walk() an iterator based walker instead of channels by @harshavardhana
• decom: Avoid objects without enough quorum by @vadmeste
• Implementing ObjectSummary API by @zveinn
• Update aistor and operator paths by @reivaj05
• Use ringbuffer in ODirect CopyAligned by @klauspost
• Add testing for periodic/delayed startup IAM routines by @taran-p
• Update the license for integration tests by @shtripat
• repl: add tests for tracked delete cleanup by @poornas
• heal: Do not reset .healing.bin after upgrade from MinIO OSS by @vadmeste
• Add itertools internal package by @donatello
• fix ReplicationStatus returned to diff by @poornas
• Fix buffered streams missing final entries by @klauspost
• Add more distributed tracing by @klauspost
• Fix helm security issue by @donatello
• xl: Ignore getattr error during migration to format.bin by @vadmeste
• hotfix: Fix Minio hotfix binary path while building docker image by @vadmeste
• Fix data race in periodic routine tests by @taran-p
• fix: marker with last entry must be only for ILM by @harshavardhana
• Align data and parity blocks to erasure set size by @allanrogerr
• simplify multi-site tests by @poornas
• Fix blank page when launching console with old assets by @reivaj05
• Fix Hadoop user-agent check for spark optimization by @poornas
• Add RPC Stream Telemetry by @klauspost
• Added integration test for CVE-2025-31489 by @shtripat
• re-implement queueStore to be more memory efficient by @harshavardhana
• Add requestNode/requestHost separately for querying by @harshavardhana
• register dummy handlers for APIs to return proper errors by @harshavardhana
• fix: re-add back the queue_dir items to be replayed properly by @harshavardhana
• Ignore error text in logonce by @klauspost
• Enabled testifylint for integration tests by @shtripat
• LIMIT 1 otherwise very large event tables cause the SELECT query to h… by @marktheunissen
• Add tests to verify prometheus access is allowed by @taran-p
• Report license validation status in logger webhook by @allanrogerr
• Updated Controller & Operator with latest code by @bexsoft
• Initial batch catalog implementation by @donatello
• Update aistor controller and operator version by @cesnietor
• Prevent random generation of restricted access keys by @taran-p
• Fix shared error buffer by @klauspost
• Revert Report license validation status in logger webhook by @allanrogerr
• tracked delete: decode dir object by @poornas
• Add New Accesskey Info and OpenID Accesskey List API endpoints by @taran-p

• re-use chunked payload buffers for signed/unsigned by @harshavardhana
• fix: healing to list, purge dangling objects by @harshavardhana
• fix: listParts crash when partNumberMarker is unexpected by @harshavardhana
• hide server info to return redacted auth_token by @harshavardhana
• ilm: Expect objects with only free versions when scanning by @krisis
• testLog: add error log when TestSiteReplicationMultiSiteWithLDAP fail… by @jiuker
• audit queue review/fixes by @zveinn

• make sure to validate signature unsigned trailer stream by @harshavardhana

• the resource leak can occur when drive Close() is called by @harshavardhana
• fix: license init in test + print goroutine count by @donatello
• More distributed tracing info+fixes by @klauspost
• fix: add more context and a racy update of deploymentID by @harshavardhana

• Move from S2 to MinLZ Compression by @klauspost

• rename express flag --proto=S3Express instead by @harshavardhana
• preserve proper nodeName and sitename in audit logging by @harshavardhana
• formatting: handle peer response with incorrect pool indices by @krisis
• Fix up transport type asserts by @klauspost
• Add support to delete all versions of an object if configured. by @poornas
• avoid repeated logging for rejected hosts in grid by @harshavardhana
• Add running go routines report at end of tests by @donatello

• Support append object behavior from S3 Express Zone by @harshavardhana
• Add OpenTelemetry distributed tracing by @klauspost

• Upgrade aistor-operator (fixes go list issues) by @ramondeklein
• fix: incorrect filePath passed into cacheGet for List() by @harshavardhana
• Fix periodic routine that clears expired token revokes. by @taran-p
• Add batch job integration test by @donatello
• fix: Cleanup heal routines when context is done by @donatello
• Add retries, errors and downtime count to replication stats by @Praveenrajmani
• Upgrade AIStor embedded console by @ramondeklein

• replace BORINGCRYPTO with Go 1.24 FIPS 140-3 module by @aead
• move batch_replication list quorum as valid config entry by @harshavardhana
• support non-blocking locking behavior for responsive APIs by @harshavardhana
• fix: move background workers as standard config by @harshavardhana
• Disallow object creation in read-only license phase by @dhananjaykrutika
• feat: Support S3 express zone requirements by @harshavardhana
• Add new API endpoint to revoke STS tokens by @taran-p

• Enable use of AISTOR console by @ramondeklein
• Update AIStor dependency to address crash by @dvaldivia
• Require license for CI testing by @allanrogerr
• Added LDAP tests by @shtripat
• make internode WriteBufferSize, ReadBufferSize configurable by @harshavardhana
• do not use ListObjects() for config history listing use Walk() by @harshavardhana
• check for errors on bitrotWriter Close() by @harshavardhana
• Update queue count for large deletes by @dhananjaykrutika
• Increase timeout for integration tests by @shtripat
• pass the appropriate error for ringbuffer by @harshavardhana
• MinIO test with huge svcaccts using go framework by @shtripat
• Add forced types and whitespace lint by @klauspost
• Fix policy detach on non-normalized LDAP users by @taran-p
• fix: setting credentials during config migrate is not necessary by @harshavardhana
• simplify batch replicate for snowball and regular use-cases by @harshavardhana
• Allow STS accounts to get info about themselves by @taran-p
• Don’t enqueue objects for expiry during listing by @krisis
• fix: tags copy when batch from remote s3 to local by @jiuker
• fix: data race in bucket event target initialization by @harshavardhana
• fix: add initial configuration and role management for console UI in standalone mode by @dvaldivia
• Move VersionPurgeStatusType into replication package by @krisis
• fix: completed batch job restarting by @donatello
• fix: return policies for k8s, assumeRoleWithCert STS accounts by @harshavardhana
• Add startup routine to check for non-normalized DNs by @taran-p
• Capture only data partitions in health report by @anjalshireesh
• Enable callhome by default by @anjalshireesh
• heal: Avoid recreating a bucket without quorum by @vadmeste
• enable zero-copy sendfile() support when enabled for READs by @harshavardhana
• fix: data race in concurrent access to failureFile by @harshavardhana
• sr: avoid bucket metadata heal stopping prematurely by @poornas
• Update SRSvcAccCreate with new type by @taran-p
• convert few _MINIO envs to standard config entries by @harshavardhana
• Skip kubernetes injected env vars by @allanrogerr
• add fadvise to expunge page-cache on Linux by @harshavardhana
• update all deps for CVEs by @harshavardhana
• do not enable callhome if server is not registered by @harshavardhana
• fix: SFTP auth bypass with no pub key in LDAP by @donatello
• All integration tests must wait for MinIO to be ready by @harshavardhana
• Add GetObjectAttributes checksum type by @klauspost
• Disassociate healtracker data and internals by @klauspost
• Batch replication tests with multiple prefixes by @shtripat
• scanner healing can skip checking for object heal on non-existent pools by @harshavardhana
• cleanup: batch job funcs using unnecessary params by @donatello
• Added multi site replication tests with LDAP by @shtripat
• ci: Disable unstable test by @klauspost
• Update multi site replication tests with LDAP by @allanrogerr
• Updated AIStor package to v0.0.3 by @bexsoft
• update integration tests dependencies by @harshavardhana
• heal: Remove empty directories during the drive scanning by @vadmeste
• s3: Fix early listing stopping when ILM is enabled by @vadmeste
• Wait for deadline till bucket replication reflected by @shtripat
• Fix prometheus authentication issue by @taran-p
• support XXH3 bitrot algorithm for low-end CPUs by @harshavardhana
• Add threshold based goroutine and memory dumps by @klauspost
• Creating an atomic time load using int64 by @zveinn
• fix: save job report in batch start handler by @donatello
• test: Add tests for checkAbandonedFolder() by @vadmeste
• fix: avoid crash in globalDriveMgr if its not initialized by @harshavardhana
• Bugfix: GRPC not starting, pass the AISTOR_CERTS_DIR env var by @pjuarezd
• decom: Ignore not found buckets by @vadmeste
• release re-used buffers ASAP instead of defer by @harshavardhana
• cleanup: Simplify renameData() and writeAllMetadataWithRevert() by @vadmeste
• save IOPs for every GetObject() on multi-pool setups by @harshavardhana
• drive heal if we have enough success, do not error setList() by @harshavardhana
• Disable network server by @ramondeklein
• make sure MRF honors the selected poolIndex, setIndex value by @harshavardhana
• Update minio/pkg, madmin-go, etc by @donatello
• Use /minio/health/cluster for readiness by @shtripat
• use single loop for all endpoints by @harshavardhana
• By default set MINIO_CALLHOME_ENABLE=off for tests by @shtripat
• Fix checksum issue during updates by @ramondeklein
• Capture license info in health diagnostics by @anjalshireesh
• make sure to skip the pool properly in HealObject() for success by @harshavardhana
• Add batch job generate + list job types handler by @donatello
• Returning errFileCorrupt to trigger healing by @zveinn
• allow locking to be at max 5 distinct nodes by @harshavardhana
• use pathJoin instead of path.Join where its relevant by @harshavardhana
• run health check before reading pool.bin by @harshavardhana
• Do not prefer a pool that has 95% of filled inodes by @vadmeste
• fix: CVE https://pkg.go.dev/vuln/GO-2025-352 by @harshavardhana
• Migrate range download test to go framework by @shtripat
• Adding timeout and entry info for dir cleanup that is timing out by @zveinn
• fix: racy update to globalLicenseInfo migrate to atomic.Pointer by @harshavardhana
• Toggle notification of REPLICA event by @allanrogerr
• read from the metadata cache when present for List() by @harshavardhana
• fix: context usage in serverPools init by @harshavardhana
• Update aistor version to v0.0.4 by @cesnietor
• fix: cleanup in test code to prevent leaks by @donatello
• format-tests: Close StorageAPI after test by @krisis
• fix: races in global variable initializations by @harshavardhana
• fix: sharedLock should allow caller to abort by @donatello
• Migrate golanglint-ci config to v2 by @donatello
• fix: cleanup in tests by @donatello
• Fix make build-dev failure by @dhananjaykrutika
• Clarify error in TemporaryAccountInfo by @taran-p
• Use updated AIStor console (incl AIStor PR #3764) by @ramondeklein
• fix: leaderlock spinning when context is cancelled by @donatello
• fix: resource setup and teardown in tests by @donatello
• remove object layer init requirement in bootstrap by @harshavardhana
• fix: passing context in config related functions by @donatello
• Migrate total writes/deletes xattrs on format.json by @krisis
• ILM: Fix evaluation of rules with NewerNoncurrentVersions by @krisis
• Use AIStor v0.0.8 by @ramondeklein

• Simplify the process of passing host names in notifications by @harshavardhana
• Fix spellcheck error by @taran-p
• Migrates tests from aistor-operator repo by @shtripat
• Make format event details queryable by @krisis
• remove cache set when reverting from xl.meta backup by @harshavardhana
• reduce bucket event workers by 50% margin by default by @harshavardhana
• fix: avoid gratuitous listing of completed batch job cleanups by @harshavardhana
• tests: Enhance and add more tests from aistor-operator by @shtripat

• Proper readiness check of KES deployment by @shtripat
• fix: applying policies for k8s STS and assumeRoleWithCertificate by @harshavardhana

• Stop persisting STS and allow automatic derivation by @taran-p

• sr: Fix healing of disabled groups by @poornas
• re-use the globally allocated static byte pool memory by @harshavardhana
• allow different binaries in a distributed setup when debugging by @harshavardhana
• MinIO SNSD should stop if drive offline by @krisis
• More detailed logging by @ramondeklein
• Fix missing authorization check for PutObjectRetentionHandler by @ramondeklein
• fix: usage of tempDir via t.TempDir() by @harshavardhana
• fix: inadvertent BatchJobCancel parameter breakage by @donatello
• Fix missing abandoned children map by @klauspost
• fix: for SNSD setups notification group is empty by @harshavardhana
• fix: maintenance check to mark health for local drives as local by @harshavardhana

• Use a pooled entry for peeking by @klauspost
• restore notification JSON with the previous format by @harshavardhana
• Use optimized crc64nvme package by @klauspost

• Update go toolchain to 1.23.6 by @donatello
• Fixing dynamic config parsing by @zveinn
• Add force option to re-add site replication config by @poornas
• batch: simplify job info format handling + other fixes by @donatello

• Allow enforcing STS to be generated and exchanged to client securely by @harshavardhana
• Migrate CVE tests here from aistor-operator by @shtripat
• Removed references of aistor-operator submodule by @shtripat
• Add motivation section to format design doc by @krisis
• Don’t generate files in current folder by @ramondeklein
• sts: allow client-provided intermediate CAs by @aead
• Update integrationtests go.mod dependencies by @JoelRuizRojas
• enable ‘docker pause/unpause’ tests by @harshavardhana
• Correct docs URL in server startup by @ravindk89
• Audit regression fix by @zveinn
• Permit backup and import of site-replicator service account. by @poornas
• change TLS session cache size to match number of nodes by @harshavardhana
• Use correct file paths fot stat by @shtripat
• update all dependencies by @harshavardhana
• Enable drive reformat test by @shtripat
• Fixing file migrations by @zveinn
• fix: multipart replication with single part objects (#20895) by @poornas
• formatting: Add sleep between retries when quorum unavailable by @krisis

• This PR improves the kafka global event queue systems by @zveinn

• Improve formatting drives in a MinIO cluster by @krisis

• Small refactor - Prepping for larger event notification system refactor by @zveinn
• do not need to add endpoint scheme for profile names by @harshavardhana
• bitrot: Do not allow write() with changing buffer length by @vadmeste
• heal: Heal an object in a suspended pool by @vadmeste
• fix: Resume healing drive after node restart by @vadmeste
• do not clone tier config for every scanner cycle() by @harshavardhana
• heal: Do not heal a fresh drive in a decommissioning pool by @vadmeste
• update deps by @harshavardhana
• change all tests to use ‘minioadmin/minioadmin’ by @harshavardhana
• kms: add MINIO_KMS_REPLICATE_KEYID option by @aead
• fix: broken diagnostics information capturing by @harshavardhana
• Base code for Bucket Event Queue + exporting Global Queue methods by @zveinn
• Send folders with no objects to heal check by @klauspost
• Another pass to further refine release note automation by @ravindk89
• healthcheck: Fix reporting drives as healing by @vadmeste
• scanner: Use timer instead of ticker to save bucket results by @vadmeste
• extend expect timeouts for non-custom reader errors by @harshavardhana
• update all dependencies by @harshavardhana
• fix: superfluous formatting message, move it to bootstrap trace by @harshavardhana
• Uniting webhook config validators into a single function by @zveinn
• update tests to use 127.0.0.1 instead of localhost by @harshavardhana
• fix: fmt-gen logging use internal logger by @harshavardhana
• Added a docker based test framework using golang by @shtripat
• Remove dependency on aistor-operator builds by @allanrogerr
• Adding adjustments to fit the new kafka packages + fixing kafka confi… by @zveinn
• fix: change force-delete to be normal bucket delete under conditions by @harshavardhana
• fix: allow less quorum to format the drives by @harshavardhana
• fix: netperf and speedtest missing self in the run by @harshavardhana
• Simplify retry in waitForFormattingDrives by @krisis
• abort healing for < 50% formatted drives per set, but with success by @harshavardhana
• do not return client disconnected unless context was not canceled by @harshavardhana
• batch: allow key rotation from sse-s3 <-> sse-kms by @poornas
• fix: Make DeleteObjects more robust to complicated delete list by @vadmeste
• do not allocate buffer without decoding in IAMImport API by @harshavardhana
• Use latest aistor-operator image and code for CVE tests by @allanrogerr
• Allow root access when site-replicator-0 is not present by @vadmeste
• Make format-events available via admin trace by @krisis
• use host instead of full URL for serverInfo, netPerf, speedTest by @harshavardhana
• site replication: fix bucket deletion syncing by @poornas
• update go deps by @harshavardhana
• Purge LDAP policy mappings for DNs not under configured base DNs by @taran-p
• provide hint when --license flag is missing by @harshavardhana
• remove unused old function by @harshavardhana
• fix: proxy requests to honor global transport by @Praveenrajmani
• pass the transport to the GetProxyEndpoints function by @Praveenrajmani
• workflow: move all workflows to go.mod version by @aead
• batch: use at least one worker for batch jobs by @aead
• Adding missing env prefix to enable variable by @zveinn
• use assert helpers in integration tests by @shtripat
• heal: Prioritize healing drive that started first in a given ES by @vadmeste
• send failure audit only in authMiddleware by @harshavardhana
• allow uncached DiskInfo() responses in drive healing by @harshavardhana
• Reorganize batch framework code by @donatello
• Multi node support (+other changes) by @ramondeklein
• Enabled integration tests for CI by @shtripat
• Redact sensitive fields from DescribeBatchJob by @harshavardhana
• Add design docs for new format algorithm by @krisis
• fix: ignore os.ErrNotExist errors and also close fd on event file by @harshavardhana

• Adding automated release notes and formatting by @ravindk89

• Run latest go based CVE tests from aistor-operator by @shtripat
• No build and cache while tests by @shtripat
• fix: bigcache rare test failure in cache.Close() by @harshavardhana
• update github.com/minio/kms-go/kes to v0.3.1 by @aead
• Added range download test scenarios by @shtripat
• Remove local setup-go action as not used by @shtripat
• More delimiter fixes (ListObjectVersions, delimiter skipping) by @klauspost
• remove nil panic from kafka logger init() by @harshavardhana

• batch: Resume resume sleep after start by @vadmeste
• allow parity count to return ‘-1’ when not initialized by @harshavardhana

• Use latest certs loaded by @shtripat
• Add OriginalAction element to ListBucketsHandler by @taran-p
• Fix data races in remove peer site by @krisis
• scanner: Avoid freeze after an ES scan failure by @vadmeste
• Add StopDelim to stop recursion when delimiter is found by @klauspost
• Fix nil ptr deref in new kafka implementation by @marktheunissen
• batch: Set last listed bucket/object in the info progress by @vadmeste

• Require the access token sub/iss to be the same as the token id by @vadmeste
• fix: endpoint string format for Netperf() by @harshavardhana
• log site replication meta load errors by @poornas
• heal: Retry healing from a list of failed objects by @vadmeste
• fix drive cache metrics names by @kanagarajkm
• add http3.0 support for all S3 API calls by @harshavardhana
• Add ’edition’ in admin info & healthinfo response by @anjalshireesh
• update deps to use new random reader source by @harshavardhana

• fix: use stashBuffer allocated for the purpose by @harshavardhana
• iam: IsAllowed to evaluate only svc claims when OpenID is enabled by @vadmeste
• Add Go runtime metrics/profile by @klauspost
• openid: Add support of getting user info regularly to update claims by @vadmeste
• update deps madmin-go/v3, pkg/v3, minio/mux by @harshavardhana

• fix: migration of old events for targets with no queue_dir set by @harshavardhana
• Remove unnecessary vars by @allanrogerr
• set parity to max if parity exceeds half of total drive count by @harshavardhana
• Add a config to move dangling objects to a quarantine folder by @vadmeste

• s3: PutObjectTagging to return 400 instead of 503 with bad XML by @vadmeste
• migrate to franz-go library from sarama by @harshavardhana
• Add resiliency tests by @allanrogerr

• update miniohq/license with broken MINIO_LICENSE parsing

• move readLock on uploadId at EOF in PutObjectPart by @harshavardhana in https://github.com/miniohq/eos/pull/258
• heal: Include more use cases of non-healable but readable objects by @vadmeste in https://github.com/miniohq/eos/pull/248
• Add grid RPC stream traces by @klauspost in https://github.com/miniohq/eos/pull/259
• update all dependencies and use latest msgp (#20768) by @harshavardhana in https://github.com/miniohq/eos/pull/261
• Implementing a new global queueing system. Refactors to audit logging, webhook logging and system logging. by @zveinn in https://github.com/miniohq/eos/pull/173

• Add certificate expiration metric (v3 only) by @ramondeklein in https://github.com/miniohq/eos/pull/234
• upgrade msgp to the latest release by @harshavardhana in https://github.com/miniohq/eos/pull/237
• fix: groups lookup performance issue with users with lots of groups by @harshavardhana in https://github.com/miniohq/eos/pull/230
• Add Cluster StorageInfo cache by @klauspost in https://github.com/miniohq/eos/pull/239
• Fixes for POST policy checks and the x-ignore implementation by @marktheunissen in https://github.com/miniohq/eos/pull/207
• Limit concurrent usage cache loading by @klauspost in https://github.com/miniohq/eos/pull/241
• Track storage reads/writes by @klauspost in https://github.com/miniohq/eos/pull/243
• fix: Privilege escalation in IAM import API by @donatello in https://github.com/miniohq/eos/pull/244
• Add license expiration metric (v3 only) by @ramondeklein in https://github.com/miniohq/eos/pull/238
• Add Full Object Checksums by @klauspost in https://github.com/miniohq/eos/pull/247
• s3: Sanitize the source object name in CopyObject handler by @vadmeste in https://github.com/miniohq/eos/pull/249
• Adds http server metrics to v3 by @klauspost in https://github.com/miniohq/eos/pull/242
• heal: Move CheckParts from single handler to streaming RPC by @vadmeste in https://github.com/miniohq/eos/pull/251
• Use smaller drive list buffer by @klauspost in https://github.com/miniohq/eos/pull/253
• re-implement Health handler to be precisely what we want by @harshavardhana in https://github.com/miniohq/eos/pull/245
• Track min/max time for windowed operations by @klauspost in https://github.com/miniohq/eos/pull/254
• Remove build and cache step for all the tests by @shtripat in https://github.com/miniohq/eos/pull/257
• Add interface for custom RPC logging by @klauspost in https://github.com/miniohq/eos/pull/255
• Add basic KMS tracing by @klauspost in https://github.com/miniohq/eos/pull/252
• Fix replication of checksum headers. by @poornas in https://github.com/miniohq/eos/pull/250

• Remove unnecessary logs while parsing quota configuration by @harshavardhana
• simplify trash folder deletion by @harshavardhana
• Do not increase proxy metrics for replication proxy not found objects by @bazko1
• Added tests for MinIO upgrade scenario by @shtripat
• Use context.Background for async tier config load by @krisis
• Re-enable resiliency mint tests, that docker pause a single node by @marktheunissen
• iam: Fix the import of site-replicator-0 by @vadmeste
• revert SlowDownRead/SlowDownWrite to SlowDown by @harshavardhana
• Add batch metrics by @allanrogerr
• Add targetArn label for bucket replication metrics by @shtripat
• Add batch jobs metrics documentation for v2 and v3 by @allanrogerr