Minio - Package Updates

[girish]

[3.6.1]

• Update minio to 2023-07-07T07-13-57Z
• Changelog
• fix: OpenID config initialization by @donatello in #17544
• fix: service account list for root user by @donatello in #17547
• fix: for FTP server driver allow implicit trust TLS by @harshavardhana in #17541
• s3: Fix a typo in per pool make/delete bucket errs calculation by @vadmeste in #17553
• fix: honor older scanner settings only if newer has not changed by @harshavardhana in #17564
• fix: upon DNS refresh() failure use previous values by @harshavardhana in #17561
• fix: create user without policy that is required by @jiuker in #17554
• fix: under FanOut API avoid repeated md5sum calculation by @harshavardhana in #17572

[girish]

[3.6.2]

• Update minio to 2023-07-11T21-29-34Z
• Changelog
• add variadic delays in peer notification retries by @harshavardhana in #17592
• Fix helm templating syntax in post-job by @aztechian in #17600
• Avoid per request URL parsing by @klauspost in #17593
• fix: set request ID in tracing context key by @donatello in #17602
• prometheus: Fix internode stats by @vadmeste in #17594

[girish]

[3.6.3]

• Update minio to 2023-07-18T17-49-40Z
• Changelog
• treat 0-byte objects to honor same quorum as delete marker by @harshavardhana in #17633
• fix: do not os.Exit(1) while writing goroutines during shutdown by @harshavardhana in #17640
• Corrected the API name for audit logging purpose by @shtripat in #17642
• chore: fix minor issues reported via staticcheck by @testwill in #17639
• Added endpoint and versions attributes to KMS details by @shtripat in #17350
• replication: Simplify mrf requeueing and add backlog handler by @poornas in #17171
• Use os.ReadFile instead of ioutil.ReadFile by @anjalshireesh in #17649

[girish]

[3.7.0]

• Update minio to 2023-07-21T21-12-44Z
• Changelog
• pick internode interface properly via globalLocalNodeName by @harshavardhana in #17680
• move bucket centric metrics to /minio/v2/metrics/bucket handlers by @harshavardhana in #17663
• fix: byHost realTime metrics API by @jiuker in #17681
• update grafana dashboard JSON with the new metrics by @harshavardhana in #17683
• fix: retry listing in decommissioning if it fails perpetually by @harshavardhana in #17682
• info: Report correct pool/set/disk indexes for offline disks by @vadmeste in #17695
• Add deletemarker_total metric by @krisis in #17689

[girish]

[3.7.1]

• Update minio to 2023-08-04T17-40-21Z
• Changelog
• return tags as part of Head/Get calls by @harshavardhana in #17635
• don't error when asked for 0-based range on empty objects by @flisk in #17708
• Revert "don't error when asked for 0-based range on empty objects (#17708) by @klauspost in #17713
• fix: Create metrics slice when necessary in getCacheMetrics() by @drivebyer in #17711

[girish]

[3.7.2]

• Update minio to 2023-08-09T23-30-22Z
• Changelog
• fix: batch replication pull must preserve versionID by @harshavardhana in #17805
• Ensure tracking 499 for canceled list objects v1, v2, versions calls by @vadmeste in #17804
• replication: set context timeout for NewMultipartUpload calls by @poornas in #17807
• batch: Strict unmarshal yaml document to avoid user made typos by @vadmeste in #17808
• fix: do not use PrefixEnabled() logic to ignore valid objects by @harshavardhana in #17677

[girish]

[3.7.3]

• Update minio to 2023-08-16T20-17-30Z
• Changelog
• Prometheus metrics for cluster_disk_offline_total has been renamed to cluster_drive_offline_total
• You may setup additional alerts on this to look for slow drives that timeout often, and also drives that return I/O errors at various intervals.
• ListObjects() continuation had a bug where in the continuation token would be ignored leading to expensive list operations when there are single prefixes with millions of objects.

[girish]

[3.7.4]

• Update minio to 2023-08-23T10-07-06Z
• Changelog
• fix: validate incoming uploadID to be base64 encoded by @harshavardhana in #17865
• do not log client canceled events by @harshavardhana in #17838
• if object is a delete marker it must skip tags filter in ILM by @harshavardhana in #17861
• ilm: Fix cleaning non current null versions by @vadmeste in #17876
• update minio/kes-go dep to v0.2.0 by @aead in #17850
• fix: replace req context by locker context by @jiuker in #17880

[girish]

[3.7.5]

• Update minio to 2023-08-29T23-07-35Z
• Changelog
• mark replication target offline if network timeouts seen by @poornas in #17907
• Retain current and upto NewerNoncurrentVersions versions by @krisis in #17909
• add missing IAM node metrics to cluster and node endpoint by @harshavardhana in #17908
• remove mTime requirement from pre-condition checks by @harshavardhana in #17916
• remove strict persistence requirements for List() .metacache objects by @harshavardhana in #17917
• fix: replace context by timeout-context from parent-context when selfSpeedTest by @jiuker in #17906

[girish]

[3.8.0]

• Update minio to 2023-08-31T15-31-16Z
• Changelog
• A major refactor of Replication metrics now provides detailed list of metrics for monitoring, alerts.
• New dashboards are introduced specifically for monitoring replication traffic between sites https://github.com/minio/minio/tree/master/docs/metrics/prometheus/grafana

[girish]

[3.8.1]

• Update minio to 2023-09-04T19-57-37Z
• Changelog
• use buffers for pathJoin, to re-use buffers. by @harshavardhana in #17960
• kafka: Do not require key when sending a message by @vadmeste in #17962
• with xlv2 format we never had to fill in checksumInfo() by @harshavardhana in #17963
• replication: remove check for empty version id by @poornas in #17964

[girish]

[3.8.2]

• Update minio to 2023-09-07T02-05-02Z
• Changelog
• metrics: fix panic in replication stats reporting by @poornas in #17979
• refactor bandwidth throttling for replication target by @harshavardhana in #17980
• use syscall.Rename() directly instead of os.Rename() by @harshavardhana in #17982
• replication resync: avoid blocking on results channel. by @poornas in #17981
• Embed file in ZIP with custom permissions by @shtripat in #17954

[girish]

[3.8.3]

• Update minio to 2023-09-16T01-01-47Z
• Changelog
• Address some significant memory issues in replication MRF when many objects are from existing object replication
• Improve Disk replacement healing from single-threaded to multi-threaded via relying on the nr_requests value on Linux
• Allow concurrent SFTP connections

[girish]

[3.8.4]

• Update minio to 2023-09-20T22-49-55Z
• Changelog
• Major memory improvements in replication hot-path reduction of 2x to 3x allocations on a busy cluster.
• SSE-KMS now uses a fast-path without doing md5sum providing a major performance boost for encrypted objects via SSE-KMS.
• Improve performance for Console UI logins.
• Fixes a regression in permission handling of LDAP user based service accounts in previous release.

[girish]

[3.8.5]

• Update minio to 2023-09-23T03-47-50Z
• Changelog
• Prometheus metrics or mc admin info would take significantly longer when audit/logger webhooks were down for some reason, performance improved when these endpoints are down makes them non-blocking. #18083
• Fix a regression while loading AssumeRoleWithCertificate credentials #18088 for more details

[girish]

[3.9.0]

• Update minio to 2023-09-27T15-22-50Z
• Changelog
• Add abandoned folder scanning to metrics by @klauspost in #18076
• Improve ListObject Compatibility by @klauspost in #18099
• fix: set scanning details locally to avoid cached values by @harshavardhana in #18092
• fix: check post policy like AWS S3 by @jiuker in #18074
• site replication: allow setting bandwidth default for bucket by @poornas in #18062

[girish]

[3.9.1]

• Update minio to 2023-09-30T07-02-29Z
• Changelog
• Fixes an issue with multipart objects with multiple versions, due to rename2() optimization being incorrectly used
• AbortMultipartUpload() is now idempotent, much like AWS S3.

[girish]

[3.9.2]

• Update minio to 2023-10-07T15-07-38Z
• Changelog
• Add tests for multipart upload overwrites on versioned buckets by @harshavardhana in #18142
• Add support for resource metrics by @anjalshireesh in #18057
• Add paramaters in Helm chart to load OIDC clientSecret from Secret Resource by @alikhtag in #17784
• Fix startup formatting by @zveinn in #18156

[girish]

[3.9.3]

• Update minio to 2023-10-16T04-13-43Z
• Changelog
• fix DeleteServiceAccount API behavior by @taran-p in #18163
• fix: upon write failure on disk journal close the file properly by @harshavardhana in #18183
• fix: O_DIRECT is on only for multi-disk setups by @donatello in #18194
• Use better host names for metric errors by @klauspost in #18188
• simplify channel send() in WalkDir() by @T-M-A in #18186
• Fix current_send_in_progress metric always being zero by @mctoohey in #18160
• avoid rebalance state for getObjectTags if any by @harshavardhana in #18197

[girish]

[3.9.4]

• Update minio to 2023-10-24T04-42-36Z
• Changelog
• choose different max_concurrent requests per drive based on HDD/NVMe by @harshavardhana in #18254
• Don't skip rules with ExpiredObjectDeleteMarker by @krisis in #18256
• Fix globalDeploymentID race by @klauspost in #18275
• Remove unused config migration by @klauspost in #18277
• Fix linter errors in Windows specific code by @klauspost in #18276
• cache DiskInfo() metrics call separately by @harshavardhana in #18270