RATATOSKRATATOSK
로그인

Longhorn

v1.11.0Storage & Data
2026년 1월 29일

CHECK 8PLAN 2OTHER 134

Longhorn v1.11.0은 V2 Data Engine 관련 기능 변화와 운영 동작 조정이 함께 포함된 기능 및 유지보수 릴리스입니다. 기존 설정과 V2 Backing Image 기능의 폐기 또는 제거, 이미지 핫픽스와 오류 수정이 반영됐습니다.

영향 확인 (8)

  • securitySPDK v25.05 CVE 문제 수정

    배포에서 SPDK v25.05를 사용하는 경우에 적용됩니다.

    SPDK v25.05에서 발생하는 CVE 문제를 수정했습니다.

  • breakingLonghorn instance manager 이미지 교체

    배포에서 Longhorn instance manager 이미지를 사용하는 경우에 적용됩니다.

    longhornio/longhorn-instance-manager:v1.11.0 이미지가 longhornio/longhorn-instance-manager:v1.11.0-hotfix-1 이미지로 교체됐습니다.

  • breakingLonghorn manager 이미지 교체

    배포에서 Longhorn manager 이미지를 사용하는 경우에 적용됩니다.

    longhornio/longhorn-manager:v1.11.0 이미지가 longhornio/longhorn-manager:v1.11.0-hotfix-1 이미지로 교체됐습니다.

  • breakingbackupstore 관련 설정 제거

    백업 저장소 관련 설정을 사용하는 경우에 적용됩니다.

    backupstore 관련 설정을 제거했습니다.

  • breaking블록 볼륨의 Trim 작업 거부

    블록 볼륨에서 Trim 작업을 실행하는 경우에 적용됩니다.

    블록 볼륨에서 Trim 작업을 거부하도록 변경했습니다.

  • breakingauto-delete-pod-when-volume-detached-unexpectedly 적용 범위 제한

    예기치 않게 볼륨을 잃을 수 있는 워크로드에 auto-delete-pod-when-volume-detached-unexpectedly를 사용하는 경우에 적용됩니다.

    auto-delete-pod-when-volume-detached-unexpectedly 설정의 적용 범위를 Kubernetes 기본 제공 워크로드로 제한했습니다.

  • breaking볼륨 사양 필드 업데이트 제한

    볼륨 사양의 spec.backingImage 또는 spec.encrypted를 업데이트하는 경우에 적용됩니다.

    두 데이터 엔진 모두에서 볼륨 사양의 spec.backingImage와 spec.encrypted를 변경할 수 없도록 했습니다.

  • breakinginstance manager CR의 폐기 필드 제거

    instance manager CR 또는 그 안의 폐기된 instances 및 instance type 필드를 사용하는 경우에 적용됩니다.

    instance manager CR에서 더 이상 사용되지 않던 instances 필드와 instance type 필드를 제거했습니다.

미리 준비 (2)

  • deprecatedV2 Data Engine의 Backing Image 기능 폐기v1.12.0에서 제거 예정

    V2 Data Engine에서 Backing Image를 사용하는 경우에 적용됩니다.

    V2 Data Engine의 Backing Image 기능이 v1.11.0에서 더 이상 사용되지 않는 기능으로 지정됐으며, v1.12.0에서 제거될 예정입니다.

  • deprecatedV2 Backing Image Feature 폐기

    V2 Backing Image Feature를 사용하는 경우에 적용됩니다.

    V2 Backing Image Feature를 더 이상 사용되지 않는 기능으로 지정했습니다.

그 외 기록된 변경 134건 전체fixes 84 · additions 25 · value changes 20 · constraints 4 · removals 1

fixes (84)

  • [BUG] Inherit namespace for longhorn-share-manager in FastFailover mode
  • [IMPROVEMENT] Fix V2 Volume CSI Clone Slowness Caused by VolumeAttachment Webhook Blocking
  • [BUG] V2 DR volume failed if backupstore is temporarily unavailable after node reboot
  • [BUG] SnapshotBack proxy request might be sent to incorrect instance-manager pod
  • [BUG] Replica rebuild, clone and restore fail, traffic being sent to HTTP proxy
  • [BUG] instance-manager on nodes that don't have hard or solid state disk DDOSing cluster DNS server with TXT query _grpc_config.localhost
  • [BUG] RWX volume causes process uninterruptible sleep
  • [BUG] Healthy replica could be deleted unexpectedly after reducing volume's number of replicas
  • [BUG] Auto balance feature may lead to volumes falling into a replica deletion-recreation loop
  • [BUG] Data locality enabled volume fails to remove an existing running replica after numberOfReplicas reduced
  • [BUG] Single replica volume could get stuck in attaching/detaching loop after the replica node rebooted
  • [BUG] v2 volume rebuild performance doesn't improve after enabling snapshot integrity
  • [BUG] Request Header Or Cookie Too Large in Web UI with OIDC auth
  • [BUG] Potential Instance Manager Client Context Leak
  • [BUG] v2 DR volume becomes faulted during incremental restoration after source volume expansion
  • [BUG] rebuildConcurrentSyncLimit field is omitted from volume.spec when value is 0
  • [BUG] Adding multiple disks to the same node concurrently may occasionally fail
  • [BUG] unknown OS condition in node CR is not properly removed during upgrade
  • [BUG] Longhorn charts does not take care timezone
  • [BUG] Pod failed to use an activated DR volume, got UNEXPECTED INCONSISTENCY; RUN fsck MANUALLY error
  • [BUG] v2 volumes do not reuse failed replicas for rebuilding as expected
  • [BUG] v2 volumes complete offline rebuilding with an extra failed replica if a node is rebooted during the rebuild
  • [BUG] Replica mode becomes empty and replica rebuilding cannot be triggered after upgrading from v1.10.1 to master-head or v1.11.0-rc1
  • [BUG] v2 volumes get stuck in Attaching/Detaching loop after node reboots
  • [BUG] Longhorn nodes may fail to recover after node reboots
  • [BUG] Missing Frontend default value when creating v2 volumes via Longhorn UI
  • [BUG] setting values are not converted to strings in Longhorn UI
  • [BUG] disk health information appears briefly
  • [BUG] encrypted v2 volume gets stuck in Attaching/Detaching loop after volume expansion
  • [BUG] Unexpected orphaned replica is created after node reboot, preventing new replica from being scheduled on that node, and blocking v2 volume from recovering to healthy state
  • [BUG] RWX volume becomes unavailable after drain node
  • [BUG] invalid memory address or nil pointer dereference
  • [BUG] share-manager excessive memory usage
  • [BUG] Encrypted Volume Cannot Be Expanded Online
  • [BUG] Backing image download gets stuck after network disconnection
  • [BUG] Can not delete the parent of volume head snapshot of a v2 volume
  • [BUG] changing of the volume controller owner caused: BUG: multiple engines detected when volume is detached
  • [BUG] mounting error is not properly handled during CSI node publish volume
  • [BUG] Flooding messages Failed to resolve sysfs path for \"/sys/class/block/root\ ... in longhorn-manager
  • [BUG] v2 volumes could fail to auto salvage after cluster restart
  • [BUG] The auo generated backing image pod name is complained by kubelet
  • [BUG] v2 volume workload FailedMount with message Staging target path /var/lib/kubelet/plugins/kubernetes.io/csi/driver.longhorn.io/xxx/globalmount is no longer valid
  • [BUG] [v1.10.0-rc1] v2 DR volume stuck Unhealthy after incremental restore with replica deletion(test_rebuild_with_inc_restoration)
  • [BUG] v2 volume could get stuck in Detaching indefinitely after node reboot
  • [Bug] A cloned volume cannot be attached to a workload
  • [BUG] Block Mode Volume Migration Stuck
  • [BUG] Replica auto balance disk pressure threshold stalled with stopped volumes
  • [BUG] short name mode is enforcing, but image name longhornio/longhorn-manager:v1.10. │ │ 0 returns ambiguous list
  • [BUG] invalid memory address or nil pointer dereference (again)
  • [BUG] Restored v2 volume gets stuck in RestoreInProgress state if backup is deleted during restoration
  • [BUG] spdk_tgt is crashed due to SIGSEGV
  • [BUG] Longhorn ignores Replica Node Level Soft Anti-Affinity when auto balance is set to best-effort
  • [BUG] SPDK NVMe synchronous calls
  • [BUG] Replicas accumulate during engine upgrade
  • [BUG] Some default settings in questions.yaml are placed incorrectly.
  • [BUG] Chart does not handle defaultSettings.taintToleration with a trailing colon
  • [BUG] Potential BackingImageManagerClient Connection and Context Leak
  • [BUG] Instance manager pod awsIAMRoleArn annotation disappearing
  • [BUG] Node block-type disk is unable to unbind after Longhorn uninstall
  • [BUG] longhorn-manager fails to start after upgrading from 1.9.2 to 1.10.0
  • [BUG][UI] When creating volume/backing image, change Data Engine will reset Number of Replicas
  • [BUG] Backup target metric is broken
  • [BUG] panic: runtime error: invalid memory address or nil pointer dereference
  • [BUG] Unable to complete uninstallation due to the remaining backuptarget
  • [BUG] NVME disk not found in v2 data engine (failed to find device for BDF)
  • [BUG] NPE error during recurring job execution
  • [BUG] v2 volume creation failed on talos nodes
  • [BUG] DR volume gets stuck in unknown state if engine image is deleted from the attached node
  • [BUG] Volume gets stuck in attaching state if engine image image is not deployed on one of nodes
  • [BUG] Rebooting the volume attached node during a v2 DR volume incremental restoration, the restoration is left incomplete and the activation has no effect
  • [BUG] Unable to detach a v2 volume after labeling disable-v2-data-engine=true
  • [BUG] Shebang refactor in scripts may cause compatibility issues
  • [BUG] Unable to re-add block-type disks by BDF after re-enable v2 data engine
  • [BUG] V2 volume stuck in volume attachment (V2 interrupt mode)
  • [BUG] Goroutine leak in instance-manager when using v2 data engine
  • [BUG] csi-provisioner silently fails to create CSIStorageCapacity if dataEngine parameter is missing
  • [BUG][v1.8.x] v2 volume stuck at attaching due to stopped replica
  • [BUG] longhorn-engine's UI panics
  • [BUG] v2 volume workload gets stuck in ContainerCreating or Unknown state with FailedMount error
  • [BUG] Volume is unable to upgrade if the number of active replicas is larger than volume.spec.numberOfReplicas
  • [BUG] UI fails to deploy when only IPv4 is enabled on nodes with v1.10.0 version
  • [BUG] v2 DR volume fails to auto-reattach when engine image missing on current node
  • [BUG] inconsistent behavior of v2 volume after labeling disable-v2-data-engine to the volume attached node and deleting the instance manager
  • [BUG] Block disk deletion fails without error message

additions (25)

  • Longhorn supports configuring UBLK performance parameters globally, per volume, or via StorageClass to improve I/O performance.
  • The V1 Data Engine now supports parallel rebuilding.
  • Longhorn improves the disk selection for the replica scheduling by introducing an intelligent balance-aware scheduling algorithm
  • Longhorn now actively monitors the physical health of the underlying disks used for storage by using S.M.A.R.T. data.
  • Users can now configure an extra network interface for the Share Manager to support complex network segmentation requirements.
  • Administrators can now use the allowedTopologies field in Longhorn StorageClasses
  • [FEATURE] Guess Linux distro from the package manager
  • [FEATURE] Provide a helm chart setting to define the managerUrl
  • [FEATURE] Add metric for last backup of a volume
  • [FEATURE] Real-time volume performance monitoring
  • [FEATURE] custom annotation/label of UI's k8s service on value.yaml of helm chart
  • [FEATURE] Make longhornctl load ublk_drv module when kernel version is 6 or newer
  • [FEATURE] Enable CSI pod anti-affinity preset update
  • [FEATURE] node storage scheduled metrics
  • [IMPROVEMENT] Introduce Concurrent Job Limit for Snapshot Operations
  • [IMPROVEMENT] Configure rolling update strategy for longhorn-manager and CSI deployments
  • [IMPROVEMENT] Add setting to disable node disk health monitoring
  • [UI][IMPROVEMENT] Awareness of when an offline replica rebuilding is triggered for an individual volume
  • [IMPROVEMENT] Add Resource Limits Configuration for Longhorn manager/instance-manager
  • [IMPROVEMENT] Add Validation Webhook to Volume Expansion When Node Disk Is Full
  • [IMPROVEMENT] Add Resource Limits Configuration for CSI Components
  • [IMPROVEMENT] Add loadBalancerClass value inside a helm chart for ui service
  • [IMPROVEMENT] Add DNS round-robin load balancing to the pool of S3 addresses
  • [IMPROVEMENT] Helm chart Multiple TLS FQDNs
  • feat(chart): Add Gateway API HTTPRoute support for Longhorn UI

value changes (20)

  • This release marks a major milestone, with the **V2 Data Engine** officially entering the **Technical Preview** stage
  • [FEATURE] Change volume operation menu button behaviour from hover to click.
  • [FEATURE] "hard" podAntiAffinity for csi-attacher/csi-provisioner/csi-resizer/csi-snapshotter
  • [IMPROVEMENT] Propagate longhorn-manager's timezone to instance-manager and CSI pods
  • [IMPROVEMENT] To improve the debugging process and UX, it would be nice that the error is recorded in the instancemanager.status.conditions.
  • [IMPROVEMENT] Avoid repeat engine restart when there are replica unavailable during migration
  • [IMPROVEMENT] Create Volume UI improvement, Automatically Filter Data Source Based on v1 or v2 Selection
  • [IMPROVEMENT] Disable the snapshot of v1 volume hashing while it is being deleted
  • [IMPROVEMENT] Ensure synchronized upgrades between longhorn-manager and instance-manager
  • [IMPROVEMENT] Use label-based state in metrics instead of numeric values
  • [UI][IMPROVEMENT] Should Not Hide the Deleted Snapshots on UI
  • [UI][IMPROVEMENT] UI shows the backing image virtual size
  • [UI][IMPROVEMENT] Show Error Message for Unschedulable Disks
  • [IMPROVEMENT] CSIStorageCapacity objects must show schedulable (allocatable) capacity
  • [IMPROVEMENT] Improve Helm Chart defaultSettings handling with automatic quoting and multi-type support
  • [TASK] Update nvme and libnvme to v2.16 and v1.16.1
  • [TASK] Revert Base Image bci-base:16.0 to bci-base:15.7
  • [TASK] Update Longhorn v1.11.0 SPDK to v25.09
  • [TASK] Bump Longhorn Component registry.suse.com/bci/bci-base to 16.0
  • has officially graduated to the **Technical Preview** stage

constraints (4)

  • Full support for the Kubernetes ReadWriteOncePod access mode has been added.
  • [IMPROVEMENT] Generalize the offline rebuilding setting for both data engines
  • [IMPROVEMENT] Check that block device is not in use before creating disk
  • [IMPROVEMENT] CRD enum values

removals (1)

  • [IMPROVEMENT] Removing executables from mirrored-longhornio-longhorn-engine image
Longhorn 스택에 추가

조치가 필요한 릴리스가 나왔을 때 주간 메일로 알려드립니다. 이번 릴리스의 보안 패치와 브레이킹 체인지 같은 것들입니다.

스택에 추가 →
공식 릴리스 노트 ↗Longhorn 릴리스 전체 →