Longhorn v1.13.0 Release Notes

The Longhorn team is excited to announce the release of Longhorn v1.13.0. This feature release brings live upgrade to the V2 Data Engine, which reached general availability in v1.12.0. V2 volumes can now stay attached while Longhorn is upgraded, as long as your cluster meets the live upgrade prerequisites.

Longhorn v1.13.0 also introduces full interrupt mode for the V2 Data Engine, volume group snapshots, age-based retention for recurring jobs, topology-constrained replica scheduling, a scheduler extender, and the longhorn-global-manager Deployment that lowers control-plane overhead in large clusters.

For terminology and background on Longhorn releases, see Releases.

Breaking Changes

Kubernetes v1.34 Minimum Version

Because the CSI external-provisioner is upgraded to v6.3.0, all clusters must be running Kubernetes v1.34 or later before installing or upgrading to Longhorn v1.13.0.

Legacy V2 Linked-Clone Volumes

Since v1.12.1, V2 linked-clone volumes created in Longhorn v1.12.0 or earlier can only be detached or deleted. To replace one, create a new linked clone from the same source volume; no data copy is required.

GitHub Issue #12552

Primary Highlights

V2 Data Engine

The V2 Data Engine became generally available in Longhorn v1.12.0. Longhorn v1.13.0 adds live upgrade and full interrupt mode, and enables CPU isolation by default for new installations.

For the expected behavior differences between V1 and V2 volumes and the feature support matrix, see V1 and V2 Volume Feature Support.

GitHub Issue #6229

Live Upgrade

Longhorn v1.13.0 supports live upgrade of V2 volumes, so the V2 instance managers can be upgraded without detaching the volumes, preventing disruption during Longhorn upgrades. For more information, see V2 Data Engine Instance Manager Upgrade.

GitHub Issue #9104

Full Interrupt Mode

In Longhorn v1.13.0, interrupt mode for the V2 Data Engine is now fully event-driven. SPDK reactors wait for I/O events instead of polling continuously, with only low-frequency background checks remaining, significantly reducing CPU usage for idle or low I/O workloads. This improves the hybrid implementation introduced in v1.10.0, which still incurred a minimal, constant CPU load when volumes were idle. As before, interrupt mode may introduce slightly higher I/O latency than polling mode under sustained high-throughput workloads.

Polling mode remains the default. To switch, set data-engine-interrupt-mode-enabled to {"v2":"true"} while no V2 volumes are attached. For more information, see Interrupt Mode Support.

GitHub Issue #11662

CPU Isolation Enabled by Default

CPU isolation keeps interrupts and other kernel background work off the CPU cores used by the V2 Data Engine. Longhorn v1.13.0 sets data-engine-cpu-isolation-enabled to {"v2":"true"} by default for new installations; clusters upgraded from v1.12.x keep their existing value and can opt in by setting it to {"v2":"true"}. CPU isolation applies only in polling mode; Longhorn skips it automatically when interrupt mode is enabled. For more information, see Data Engine CPU Isolation Enabled.

GitHub Issue #13724, GitHub Issue #13973

Fast Volume Cloning

Since v1.12.1, fast volume cloning for the V2 Data Engine has used a linked-clone architecture: a linked-clone volume shares data blocks with its source instead of copying them. One source volume can back multiple linked clones. Linked-clone volumes support most operations available to regular volumes, including snapshots, expansion, replica rebuilding, backup, and use as the source of nested linked clones. Starting with v1.13.0, V2 linked-clone volumes also support restore operations. For more information, see Volume Clone Support.

GitHub Issue #12552

Storage Sharding (Experimental)

Introduced in v1.12.1, storage sharding is an experimental V2 Data Engine feature. Instead of keeping a full copy of the volume on every replica, it splits written data into data and parity chunks and spreads them across nodes. A volume can then be larger than any single disk or node, and tolerate the same number of failures with less disk space.

Sharded volumes do not support backup and restore, volume cloning, backing images, disaster recovery (DR) volumes, or live migration. This feature is for evaluation and testing only and is not recommended for production. For more information, see Sharding with Erasure Coding.

GitHub Issue #1061

SPDK iobuf Pool Size Configuration

Introduced in v1.12.1, the data-engine-iobuf-large-pool-size and data-engine-iobuf-small-pool-size settings control the size of the SPDK buffer pools used by the V2 Data Engine. Larger pools help avoid running out of buffers under high-queue-depth workloads. Because the pools are sized when SPDK starts, changing either setting recreates V2 instance manager pods that have no running engines or replicas.

GitHub Issue #13322, GitHub Issue #13674

Snapshots and Backups

Volume Group Snapshot Support

Longhorn v1.13.0 makes it easier to protect related volumes by letting you snapshot a group of volumes with a single request from the Longhorn UI, kubectl, or Kubernetes VolumeGroupSnapshot objects through CSI. The CSI path, which is disabled by default, also supports group backups.

For more information, see Create a Snapshot Group and Enable CSI Volume Group Snapshot Support.

GitHub Issue #13349

Age-Based Retention for Recurring Jobs

Longhorn v1.13.0 adds an age-based retention policy for snapshot, backup, and system backup recurring jobs, letting you retain data based on its age instead of the number of items. Set retentionPolicy to age-based and retainAge to a duration such as 720h; each run deletes snapshots or backups older than the specified age. The count-based policy remains the default, and existing recurring jobs are unchanged after the upgrade.

GitHub Issue #12060

Smarter Scheduling

Volume Topology Constraint

Longhorn v1.13.0 adds the volumeTopology StorageClass parameter (any, zonal, or regional) to keep a volume's replicas within the zone or region where it was provisioned, including during rebuilds and replica count changes, so replicas stay close to the workload. Previously, zone labels only spread replicas across zones; a rebuild could still place a replica in a different zone from the workload. For more information, see Topology-Aware Provisioning.

GitHub Issue #13493

Scheduler Extender

Longhorn v1.13.0 improves pod scheduling by allowing kube-scheduler to check actual Longhorn disk capacity and, when possible, place a restarted pod on the node that already holds all of its replicas. This helps avoid scheduling decisions based on stale CSIStorageCapacity data during bursts of pod creation and improves data locality, especially with best-effort data locality. The scheduler extender runs inside longhorn-manager and requires a kube-scheduler configuration change, which is not possible on managed Kubernetes services such as GKE and EKS.

GitHub Issue #12591

Better Operations

Longhorn Global Manager

Longhorn v1.13.0 reduces kube-apiserver load and longhorn-manager memory usage in large clusters by moving the cluster-wide pod and PersistentVolume controllers from the longhorn-manager DaemonSet to a new longhorn-global-manager Deployment. The Deployment is created with three replicas by default during installation and upgrade. Before upgrading, make sure at least one of its pods can be scheduled. For more information, see Upgrading Longhorn Manager.

GitHub Issue #13059

Security Hardening

Internal Network Policies

Since v1.12.1, Longhorn creates ingress NetworkPolicy resources for its internal components by default. They take effect only when the CNI plugin enforces NetworkPolicy. Longhorn v1.12.2 resolves the CNI compatibility issues found in v1.12.1 and adds three Helm values:

• networkPolicies.v1DataEngineInitiatorSourceCIDRs and networkPolicies.recoveryBackendAdditionalIngressPorts allow traffic that the v1.12.1 policies blocked.
• networkPolicies.metricsScrapeSources lets Prometheus and other scrapers reach longhorn-manager metrics on TCP port 9500.

For more information, see Internal Network Policies.

GitHub Issue #13438, GitHub Issue #13802, GitHub Issue #13740, GitHub Issue #13947

Instance Manager gRPC mTLS Coverage

Before v1.12.1, when the longhorn-grpc-tls secret was configured, mutual TLS (mTLS) covered only the instance manager's instance and proxy gRPC services; the disk and SPDK services still accepted plaintext connections. Since v1.12.1, mTLS covers all instance manager gRPC services, so every gRPC port requires a valid client certificate when the secret is configured.

GitHub Issue #7787, GitHub Issue #13212

Dedicated CSI Service Account

Longhorn v1.13.0 runs the CSI controller sidecars under a dedicated longhorn-csi-service-account instead of the shared longhorn-service-account. For compatibility with existing Secret references, the new service account still receives cluster-wide get access to Secrets by default. You can turn this off with the csi.allowControllerSecretAccess Helm value after removing csi.storage.k8s.io/provisioner-secret-* parameters from Longhorn StorageClasses. For more information, see Optional Restriction of CSI Controller Secret Access.

GitHub Issue #14020

Critical Stability Fixes

Linked-Clone Backup Restore

Longhorn v1.13.0 fixes a V2 linked-clone backup restore issue that could produce a corrupted volume when the source volume or the snapshot the clone was created from no longer existed. Backups of V2 linked-clone volumes now record the source volume and snapshot, and a restore fails with an error if either is missing.

GitHub Issue #13714

CSI Volume Clone with Strict-Local Data Locality

Longhorn v1.13.0 fixes a CSI volume clone issue where cloning a volume with dataLocality: strict-local could fail with hard affinity cannot be satisfied. Longhorn now picks a node that matches the volume's nodeSelector and diskSelector for the clone.

GitHub Issue #12792

Longhorn Node Removal After Kubernetes Node Deletion

Longhorn v1.13.0 fixes a node cleanup issue where a Longhorn node could not be removed after its Kubernetes node had been deleted without eviction. You can now disable scheduling on the node, remove its remaining replicas and engines, and then delete it. Evicting a node before deleting it from the cluster is still the recommended procedure. For more information, see Graceful Node Removal.

GitHub Issue #13494

Backing Image Copies on IPv6 Clusters

Longhorn v1.13.0 fixes a backing image issue where copies were always transferred over IPv4, so on IPv6 single-stack and IPv6-first dual-stack clusters a backing image stayed at one copy. Backing images are now copied over the cluster's IP family and the storage network.

GitHub Issue #13864

Installation

You can install Longhorn using a variety of tools, including Rancher, kubectl, and Helm. For more information about installation methods and requirements, see Quick Installation in the Longhorn documentation.

Upgrade

Longhorn only allows upgrades from supported versions. For more information about upgrade paths and procedures, see Upgrade in the Longhorn documentation.

Automated pre-upgrade checks do not cover all scenarios. Before upgrading, review the manual checks in Important Notes, and for V2 volumes confirm the live upgrade prerequisites or detach them and ensure their replicas are stopped before upgrading.

Post-Release Known Issues

For information about issues identified after this release, see Release-Known-Issues.

Resolved Issues in this release

Highlight

• [FEATURE] Support v2 Data Engine (GA) 6229 - @derekbit
• [FEATURE] V2 Volume Supports Live Upgrade 9104 - @yangchiu @davidcheng0922
• [FEATURE] VolumeGroupSnapshot support 13349 - @c3y1huang @chriscchien
• [FEATURE] Full Interrupt Mode for V2 Data Engine 11662 - @au2001 @chriscchien
• [FEATURE] V2 Data Engine Fast Cloning 12552 - @shuo-wu @roger-ryao
• [FEATURE] V2 Data Engine Sharding - Experimental 1061 - @c3y1huang @chriscchien

Feature

• [FEATURE] Support Age-Based Retention for Recurring Jobs 12060 - @mantissahz @roger-ryao
• [UI][FEATURE] V2 Data Engine Fast Cloning 13671 - @shuo-wu @roger-ryao
• [FEATURE] Backing image supports IPv6 and dual-stack IP families 13864 - @COLDTURNIP @yangchiu
• [UI][FEATURE] VolumeGroupSnapshot support 13707 - @a110605 @chriscchien
• [FEATURE] Allow configuring Gateway API filters 12976 - @apoorvajagtap @chriscchien
• [UI][FEATURE] Support Age-Based Retention for Recurring Jobs 13669 - @yangchiu @a110605 @houhoucoop
• [FEATURE] TLS gRPC API for instance manager 13212 - @COLDTURNIP

Improvement

• [IMPROVEMENT] Restrict the Secret access of CSI components 14020 - @COLDTURNIP @roger-ryao
• [IMPROVEMENT] [CLI] preflight checks on Elemental3-based systems fails 13949 - @rdoxenham @yangchiu
• [IMPROVEMENT] Enable CPU Isolation for V2 Data Engine 13724 - @derekbit @roger-ryao @Copilot
• [IMPROVEMENT] Introduce longhorn-global-manager Deployment for cluster-wide controllers 13059 - @hookak @chriscchien
• [IMPROVEMENT] backup target availability status metric 13980 - @yangchiu @eklatzer
• [IMPROVEMENT] Disable CPU Isolation When V2 Data Engine Uses Interruption Mode 13973 - @derekbit @chriscchien
• [IMPROVEMENT] Add longhorn-cli command to check if the user's environment supports IOMMU 12076 - @yangchiu @apoorvajagtap
• [IMPROVEMENT] Make ignoreSigningHeaders configurable — SigV4 Accept-Encoding breaks every S3 backup target behind a header-rewriting proxy, not just GCS 13756 - @kinorai @roger-ryao
• [IMPROVEMENT] Longhorn Scheduler Extender 12591 - @bachmanity1
• [IMPROVEMENT] v2 data engine: gate the replica host-ACL restriction on cluster-wide instance manager version 13699 - @hookak
• [IMPROVEMENT] CSI ignores EC StorageClass parameters (silent replicated fallback) 13792 - @yangchiu @apoorvajagtap
• [IMPROVEMENT][Helm] Allow setting spec.trafficDistribution 13814 - @mydoomfr @yangchiu
• [IMPROVEMENT] (chart) Precedence of image...registry 12611 - @bachmanity1 @roger-ryao
• [IMPROVEMENT] Add metrics to collect information about V2 data engine usage 12941 - @derekbit @chriscchien @Copilot
• [IMPROVEMENT] Allow configuring SPDK iobuf large pool size 13322 - @chriscchien @bachmanity1
• [IMPROVEMENT] Truncate the attached-to pod list in the volume table for volumes with many pods 13851 - @HypeMC @chriscchien
• [IMPROVEMENT] Add a pod filter to the Workload/Pod detail popup 13852 - @HypeMC @chriscchien
• [IMPROVEMENT] Every longhorn-manager pod builds its DataStore and informer caches twice 13782 - @hookak @roger-ryao
• [IMPROVEMENT] Expose topologySpreadConstraints for longhorn-ui in the Helm chart 13731 - @yangchiu @seanblake
• [IMPROVEMENT] Respect volume topology in replica scheduling (opt-in zonal/regional volume placement) 13493 - @yangchiu @hookak
• [IMPROVEMENT] longhorn-manager watches Leases cluster-wide but reads only those in the Longhorn namespace 13786 - @hookak @chriscchien
• [IMPROVEMENT] longhorn-manager caches every Secret in the cluster in order to watch one 13726 - @yangchiu @hookak
• [IMPROVEMENT] v2 volumes: allow configuring the NVMe-TCP initiator I/O queue count (nr-io-queues) 13706 - @hookak @chriscchien
• [IMPROVEMENT] Allow configuring SPDK iobuf small pool size 13674 - @yangchiu @hookak
• [IMPROVEMENT] Increase kbench fio numjobs to higher value 9100 - @derekbit
• [IMPROVEMENT] Simplify kbench Metrics to Focus on Key Performance Indicators 9460 - @derekbit
• [IMPROVEMENT] go-common-libs: kill the child process when command execution times out 13619 - @hookak
• [IMPROVEMENT] always setup NetworkPolicy for the internal communication 13438 - @COLDTURNIP @roger-ryao
• [IMPROVEMENT] Steer host RPS away from SPDK reactor cores 13483 - @bachmanity1 @roger-ryao
• [IMPROVEMENT] updateBackupCompressionMethod may write the Volume even when the method is unchanged 13460 - @yangchiu @ChaoHuang2018
• [IMPROVEMENT] Improving error transparency for volume attachment failure 9968 - @derekbit @chriscchien @Copilot
• [IMPROVEMENT] Add usage metrics for volume size 11521 - @derekbit @chriscchien @Copilot
• [IMPROVEMENT] V2 volume write I/O stalls(~10s) when a replica is removed during migration 13309 - @hookak @chriscchien
• [IMPROVEMENT] Support mTLS encrypted communication for remaining gRPC services in instance manager 7787 - @COLDTURNIP @yangchiu
• [IMPROVEMENT] Add metrics to collect information about LONGHORN_DISTRO 13252 - @derekbit @chriscchien
• [IMPROVEMENT] Exporting a volume from a single replica via crictl command 7711 - @c3y1huang

Bug

• [BUG] Runtime container images contain development (-devel) packages 13479 - @yangchiu @benispeti
• [BUG] DR replicas break 13685 - @derekbit @chriscchien
• [BUG] Longhorn may not reject expansion requests when the volume clone is in-progress 14065 - @yangchiu @shuo-wu
• [BUG] Request Schema in Upgrade Responder is Outdate 14027 - @yangchiu @roger-ryao
• [BUG] Test case test_csi_block_volume_online_expansion fails with 'assert md5sum: == ...' on v2 volume (AMD64 only) 14045 - @derekbit @roger-ryao
• [BUG] Disconnect Volume Node Network For More Than Pod Eviction Timeout While Workload Heavy Writing With RWX Fast Failover Disabled fails on v2 data engine 13913 - @derekbit @shuo-wu @chriscchien
• [BUG] deleting nodes.longhorn.io failed due to KubernetesNodeGone 13494 - @COLDTURNIP @yangchiu
• [BUG] test test_single_replica_failed_during_engine_start fails 13861 - @davidcheng0922
• [BUG] Cloned volume stuck "not ready for workloads" after the clone finishes (v1.12.0, v1 engine) 13335 - @roger-ryao @somanchi004-code
• [BUG] SystemRestore job ignores the registry-secret setting 13968 - @bachmanity1 @roger-ryao
• [BUG] After host restart one node no longer abe to start any replica 13733 - @roger-ryao
• [BUG] test_engine_image_not_fully_deployed_perform_dr_restoring_expanding_volume fails on v1.13.x-head 13974 - @shuo-wu @chriscchien
• [BUG] v2 workload IO gets stuck in Continuous IO Test 13912 - @yangchiu @c3y1huang
• [BUG] Test case Test V2 Volume Engine Live Switchover may fail on v1.13.x-head 13951 - @yangchiu @davidcheng0922
• [BUG] Terminal workload pods cause NodePublishVolume to abort with "no Pending workload pods" 13723 - @yangchiu @guolxingxing
• [BUG] v2 volumes might get stuck in deleting state 13585 - @yangchiu @davidcheng0922 @sushant-suse
• [UI][BUG] The restore of V2 Linked Clone volumes misses src volume handling 13922 - @shuo-wu @roger-ryao
• [BUG] The restore of V2 Linked Clone volumes misses src volume handling 13714 - @shuo-wu @roger-ryao
• [BUG] Sharded/EC volume: writes stall to ~0 and wedge the SPDK target on arm64/AIO, blocking volume deletion; plus CSI drops EC StorageClass params 13789 - @c3y1huang @chriscchien
• [BUG] Regression test fails with v2 BDF block disks 13939 - @yangchiu @davidcheng0922 @sushant-suse
• [BUG] v2 volume gets stuck in Attaching in an IPv6 storage network environment 13933 - @yangchiu @derekbit
• [BUG] CSI volume cloning doesn't respect node-/diskSelector 12792 - @roger-ryao @carterli0407-cell
• [BUG] [v1.13.0-rc1] V2 volume DD easy to stuck in V2 interrupt mode 13937 - @au2001 @roger-ryao
• [BUG] v1.12.1 default NetworkPolicies miss legitimate non-pod-labeled and host-originated callers, silently breaking CSI attach/detach 13802 - @COLDTURNIP @roger-ryao
• [BUG] Replica rebuiliding may fail with Input/output error on v2 volumes 13189 - @derekbit @chriscchien
• [BUG] Test case Test DR Volume Live Upgrade And Rebuild fails 13911 - @yangchiu @derekbit
• [BUG] Creating default disks with createDefaultDiskLabeledNodes setting and create-default-disk label doesn't work with BDF 13491 - @yangchiu @davidcheng0922
• [BUG] Test case test_node_default_disk_labeled fails 13910 - @roger-ryao
• [BUG] invalid block disks should not block Longhorn system uninstallation 13572 - @apoorvajagtap @chriscchien
• [BUG] The manifests in the /deploy directory of the longhorn/cli repository do not work due to an error in the command: field. 13514 - @chattytak @roger-ryao
• [BUG] V2 replica rebuild destination failure faults the entire engine 13673 - @christophersherman @roger-ryao
• [BUG] v1 engine never starts: a stale stopped process record makes createInstance a silent no-op, volume stuck in attaching forever 13687 - @derekbit @roger-ryao
• [BUG] SPDK initialization fails on Talos OS with "nsenter: operation not permitted" 12709 - @c3y1huang @roger-ryao @sushant-suse
• [BUG] V2 expansion can report success while the engine remains at the old size 13379 - @davidcheng0922 @chriscchien
• [BUG] Recurring trim job fails with deadlock 13416 - @c3y1huang @roger-ryao
• [BUG] Kernel Workqueue Lockup and Unstable RKE2 Service After Enabling LH V2 in Harvester 13417 - @derekbit @chriscchien
• [BUG] Failed to add v2 block disk with virtio-scsi BDF path 13474 - @chriscchien @carterli0407-cell
• [BUG] Longhorn 1.12.0: AWS chunked encoding not supported with OCI S3 buckets 13477 - @derekbit @mantissahz @roger-ryao
• [BUG] Host OS nvmf-autoconnect connects kernel initiators to v2 replica subsystems, stalling volume attach/detach for minutes 13651 - @hookak @chriscchien
• [BUG] Adding V2 disk using the "Default Data Path" /dev/disk/by-id/scsi-* path fails 13557 - @davidcheng0922 @roger-ryao
• [BUG] backupvolume.longhorn.io not updated on cluster with nodes drained 13859 - @derekbit @chriscchien
• [BUG] Volume expansion snapshot CR name expand- collides across volumes 13386 - @derekbit @chriscchien
• [BUG] System managed components don't respect System Managed Components Node Selector 12834 - @COLDTURNIP @yangchiu
• [BUG] Recurring-job pods report success after startup or volume execution errors 13587 - @christophersherman @roger-ryao
• [BUG] backupvolume.longhorn.io not created on cluster have nodes drained 13775 - @derekbit @chriscchien
• [BUG] Backup not started on cluster have nodes drained 12562 - @chriscchien @carterli0407-cell
• [BUG] Test case Power Off Replica Node Should Not Rebuild New Replica On Same Node fails 13705 - @yangchiu @derekbit
• [BUG] v1 replica rebuild fails after a single gRPC TCP keepalive is lost 13703 - @dnesting @chriscchien
• [BUG] v1.12.1 chart: new default-on networkPolicies.restrictInternalTraffic silently breaks Prometheus scraping of longhorn-manager (metrics/alerting blackout on patch upgrade) 13740 - @COLDTURNIP @roger-ryao
• [BUG] Longhorn UI is not running with PID 1 13680 - @yangchiu @xandau
• [BUG][v1.6.2-rc2] Volume failed to recover after nodes reboot, pod failed to mount the volume with Input/output error 8587 - @shuo-wu @roger-ryao
• [BUG] Snapshot CR is missing for backup CR of recurring job 10184 - @yangchiu @christophersherman
• [BUG] V1 volumes not rebuilding after cluster shutdown 13571 - @COLDTURNIP @yangchiu
• [BUG] Incorrect Web Link in GUI 13352 - @yangchiu @sushant-suse
• [BUG] V2 backup/snapshot can leave NVMe/TCP frontend or dm device stale, causing pod EIO on attached volumes 13331 - @davidcheng0922 @chriscchien
• [BUG] Longhorn Helm Chart NetworkPolicies do not honor new RKE2 "rke2-traefik" ingress controller 13653 - @COLDTURNIP @roger-ryao
• [BUG] v2 volume repeated replica reuse failure 13315 - @shuo-wu @chriscchien
• [BUG] Longhorn may try to attach volumes to a node without valid IM pod during the clone 13639 - @yangchiu @shuo-wu
• [BUG] Test case test_volume_scheduling_failure fails on v2 volumes 13655 - @yangchiu @c3y1huang
• [BUG] Backup Listing With More Than 1000 Backups fails on v2 volume due to an empty replica address in the backup status 13611 - @COLDTURNIP @chriscchien
• [BUG] Interrupt Mode V2 stuck high CPU after workload completed 12066 - @c3y1huang
• [BUG] Encrypted V2 volume size is 16MB short of the claimed size 13163 - @mantissahz @roger-ryao
• [BUG] Transient SPDK lvol metadata failure can permanently fault a healthy v2 replica 13541 - @christophersherman @roger-ryao
• [BUG] CSI pods do not respect anti-affinity preset update 13546 - @chriscchien @carterli0407-cell
• [BUG] V2 encrypted volume keeps switching between Attaching and Detaching state after expand operation 13561 - @mantissahz @roger-ryao
• [BUG] Test case Recurring Job Pod Should Not Crash fails 13567 - @yangchiu @c3y1huang
• [BUG] Fail to restore a v2 volume from a full backup if a previous backup is corrupted 13526 - @yangchiu @derekbit
• [BUG] Error logs in longhorn-uninstall job 13547 - @yangchiu @c3y1huang
• [BUG] V2 Data Engine: UBLK fails with EINVAL on Linux kernel 6.17.0 11977 - @chriscchien @carterli0407-cell
• [BUG] csi.ReplicaCount Helm values silently no-op on existing csi- deployments (only applied at first creation) 13461 - @sebastiangaiser @roger-ryao
• [BUG] v2 volume may crash again after the auto reattachment 13314 - @shuo-wu @roger-ryao
• [BUG] V2 Encrypted Volume Restore Fails 13363 - @mantissahz @roger-ryao
• [BUG] (chart) ArgoCD OutOfSync when using Gateway API 13340 - @yangchiu @SalvoRusso8
• [BUG] Migration Engine Can Be Unexpectedly Deleted If the Target Node Is Still in Readiness Transition 13109 - @COLDTURNIP @yangchiu
• [BUG] volume expansion stuck 13334 - @shuo-wu @roger-ryao
• [BUG] expanding the volume fails 13355 - @AoRuiAC @chriscchien
• [BUG] Test case test_rwx_delete_share_manager_pod fails because it's unable to find the exported volume in share manager pod after it's deleted and restarted 13221 - @davidcheng0922 @roger-ryao
• [BUG] System Backup RecurringJob retention prunes newest CR — sorts by Status.CreatedAt (zero for Error/racing CRs) 13203 - @issmirnov @roger-ryao
• [BUG] CSI components may have 0 running replica during upgrade 13105 - @yangchiu @carterli0407-cell
• [BUG] Test case test_best_effort_data_locality fails because there is no replica for the created volume 13222 - @yangchiu @carterli0407-cell
• [BUG] spdk interrupt mode value is missing in chart/values.yaml 13266 - @yangchiu @kema-dev
• [BUG] Misconfigured and deleted Talos node causes Driver Deployer to constantly restart 12159 - @c3y1huang

Resilience

• [BUG] V2 instance-manager liveness probe errors ('test: -eq: unary operator expected') -> self-kill -> node-wide replica fault under rebuild load 13957 - @yangchiu @mantissahz @Copilot
• [DOC] Enhance Longhorn docs about Instance Manager 13197 - @Felipalds @chriscchien
• [BUG] V2 Instance Manager panics in backupstore when S3 volume.cfg is missing, taking all node volumes offline 13790 - @mantissahz @chriscchien

Stability

• [BUG] V2 backup readers race snapshot cleanup and panic the instance manager 14017 - @christophersherman @roger-ryao
• [BUG] Recurring snapshot-cleanup aborts the entire run when any volume's replica is rebuilding 13623 - @derekbit @chriscchien

Misc

• [DOC] the upgrade behavior for volumes differs between the v1 and v2 data engines 14023 - @mantissahz @chriscchien
• [IMPROVEMENT] Chart value to allow metrics scrapers through the longhorn-manager NetworkPolicy 13947 - @alekc @yangchiu
• [IMPROVEMENT] Help large replicas on full disks reclaim the space and squash all data into the heads 13636 - @apoorvajagtap @roger-ryao
• [DOC] Remove backing image dependency from v2 BI backup guide 13858 - @COLDTURNIP @chriscchien
• [BUG] Longhorn block disk remains Ready=False and Schedulable=False when NVMe device is bound to vfio-pci 13893 - @derekbit @chriscchien
• [DOC] Add Data Path Diagrams for V1 and V2 Data Engines 13570 - @derekbit @chriscchien
• [TASK] Clarify the share manager NFS server dual-stack compatibility 13900 - @COLDTURNIP
• [DOC] Broken image URL for OpenShift OAuth proxy 13339 - @geragio @roger-ryao
• [DOC] Create a Rancher Chart Generation and Release Wiki Page 13346 - @carterli0407-cell
• [TASK] document to cleanup orphaned backing image copies manually 13523 - @COLDTURNIP @chriscchien
• [BUG] CSI method-level request logs bypass secret sanitization 13613 - @anthonymartin @yangchiu
• [DOC] Improve contributing documentation 10814 - @derekbit
• [DOC] NetworkPolicy setup guidance 13591 - @COLDTURNIP @roger-ryao
• [DOC] Chart values.yaml still refers to Data Engine V2 as experimental 13614 - @sushant-suse
• [TASK] fix longhorn-instance-manager lint validation failure 13313 - @COLDTURNIP @roger-ryao
• [IMPROVEMENT] Add opt-in PodDisruptionBudget for the Longhorn UI Deployment 13466 - @yangchiu @somaz94
• [DOC] Update the minimum Kubernetes version requirement to v1.34 13576 - @derekbit @roger-ryao
• [BUG] v2 volume stuck attaching with Storage Network enabled because the EngineFrontend target uses the engine pod IP instead of StorageIP 13351 - @yangchiu @c3y1huang
• [FEATURE] Support Kubernetes CPU Manager for Longhorn V2 instance-manager SPDK CPU assignment 13248 - @yangchiu @mantissahz
• [REFACTOR] Refactor longhorn-spdk-engine codes 12491 - @derekbit
• [TASK] Remove downstream security patches for CSI sidecar images 13329 - @derekbit

Contributors

• @98jan
• @AoRuiAC
• @COLDTURNIP
• @ChaoHuang2018
• @DrJosh9000
• @Felipalds
• @HypeMC
• @IggyGG
• @Martin-Weiss
• @NRCan-LGariepy
• @PhanLe1010
• @SalvoRusso8
• @WebberHuang1118
• @Xeboc
• @a110605
• @abonillabeeche
• @alekc
• @anthonymartin
• @antoinemichea
• @apoorvajagtap
• @au2001
• @bachmanity1
• @benispeti
• @boomam
• @brandboat
• @c3y1huang
• @carterli0407-cell
• @chattytak
• @chriscchien
• @christophersherman
• @cosmo-wang
• @daftu
• @darkpixel
• @davidcheng0922
• @derekbit
• @divya-mohan0209
• @dnesting
• @eklatzer
• @foobardmr
• @geragio
• @guolxingxing
• @hookak
• @houhoucoop
• @innobead
• @issmirnov
• @jabrown93
• @jkempson
• @jzaehrin
• @kema-dev
• @kinorai
• @kolonelkrazy
• @kondanta
• @lictw
• @mantissahz
• @mydoomfr
• @rd-eng
• @rdoxenham
• @roger-ryao
• @seanblake
• @sebastiangaiser
• @shuo-wu
• @somanchi004-code
• @somaz94
• @steled
• @sushant-suse
• @tarlomitico
• @tvanderka
• @vanchaxy
• @xandau
• @yangchiu
• @yasker
• @yhamouda
• @zauguin
• @rebeccazzzz
• @forbesguthrie
• @asettle