Container Service for Kubernetes:csi-provisioner

• Deletion protection for disks is supported.
• Anolis OS is supported.
• Statically provisioned OSS volumes can be mounted to ACK Lingjun managed clusters.
• Hot update is supported by the CSI caching client.
• The performance of continuous integration and continuous delivery (CI/CD) pipelines and data distribution is improved.

• csi-plugin can be deployed on ECS Bare Metal instances that run the Kylin operating system.
• Multiple performance levels (PLs) are supported.
• The initialization logic is optimized.

• Kubernetes 1.24 is supported by csi-plugin.
• OSS bucket lifecycle management is supported by Container Network File System (CNFS).
• Observability is supported by the caching feature of CNFS.

Kubernetes 1.24 is supported by csi-plugin.

• I/O operation monitoring is supported when OSS volumes are mounted to pods.
• ossfs is optimized for downloads of large files. The default block size for block downloads is changed from 5 MB to 20 MB. The number of concurrent download threads is changed from 5 to 20.
• The protocol for API calls of all Alibaba Cloud services is changed from HTTP to HTTPS.
• The ConfigMap used to configure canary release settings is deleted from Cloud Parallel File Storage (CPFS) 2.0 and alinas.
• Snapshots can be created for enhanced SSDs (ESSDs) during disk expansion.

• Disks can be mounted to Edge Node Service (ENS) nodes.
• Dynamically and statically provisioned volumes that use Cloud Parallel File Storage (CPFS) 2.0 are supported.
• The caching feature provided by the Container Network File System (CNFS) client is supported.
• Statically provisioned alinas volumes are supported and TLS can be used for encrypted transmission.
• Statically provisioned Object Storage Service (OSS) volumes can be mounted to JindoFS.
• OSSFS 1.87 is released. The issue that OSS volumes fail to be mounted due to duplicate entries in concurrent processing scenarios is fixed.

• The issue that ossfs responds slowly to the ls command on Alibaba Cloud Linux 3 is fixed.
• The features of the Container Storage Interface (CSI) plug-in in Kubernetes 1.22 are supported. Kubernetes clusters of Kubernetes 1.22 are supported.
• Before you delete a subdirectory in a NAS file system that is mounted by using CNFS, the quota limit of CNFS is removed.
• The IOPS and throughput of volumes that are managed by Logical Volume Manager (LVM) can be limited by using cgroups.
• The capacity of Non-Volatile Memory Express (NVMe) SSDs can be expanded.
• Vulnerabilities in the CentOS 7 base image are patched.

• TLS authentication is supported by LVM.
• csi-provisioner must be updated before csi-plugin when you update CSI.
• If the quota feature is enabled for a NAS file system and the NAS file system is mounted in subpath mode, you must disable the quota feature before you can delete the persistent volume (PV) that is used to mount the NAS file system.

• The issue that ossfs cannot synchronize data in real time is fixed.
• The issue that NAS volumes fail to be mounted is fixed.

• The sharepath feature of NAS is supported.
• The issue that the component is frequently restarted due to the unstable etcd is fixed.
• The ARM64 and AMD64 architectures are supported.

• The time parameter for the recycle bin is added to CNFS.
• The apiVersion of CNFS is changed from v1alpha1 to v1beta1.
• The issue that ossfs cannot synchronize data in real time is fixed.
• By default, the forcibly detach disk option is disabled.

• The issue that NAS file systems cannot be expanded by using CNFS is fixed.
• Object Storage Service (OSS) buckets can be mounted to nodes that are deployed by using the Alibaba Cloud Linux 3 image.

• The volumeCapacity field is deleted from NAS volume configurations. The allowVolumeExpansion field is used to specify whether to enable the quota feature.
• The selflink field is deleted from NAS volume configurations.

• Disk partitions can be mounted.
• Disk partitions can be expanded.

• Kubernetes 1.20 is supported. The metadata.selflink field is deleted.
• The tag of the cluster ID is automatically added to disks.
• NAS volumes can be expanded within the quota limit.

• Database File System (DBFS) volumes are supported.
• By default, volume monitoring is enabled.
• Local volumes of the QuotaPath type are supported.
• The VolumeSnapshot List feature is supported.
• Quota groups are supported by NAS volumes.
• Custom disk types are supported.

• The deployment template is updated to merge drivers into one container.
• The issue that subdirectories fail to be created in Extreme NAS file systems is fixed.
• Kubernetes 1.18 is supported.
• Labels can be added to NAS volumes when you create NAS volumes.

• The issue that snapshots cannot be created from disks is fixed.
• The issue that dynamic provisioning of NAS volumes fails due to residual data is fixed.
• The check logic of BDF nodes when csi-provisioner is started is fixed.
• The use of UUID to obtain device paths is no longer supported.

• Elastic Block Storage (EBS) snapshots are supported. You can use EBS snapshots to restore data to a beta version.
• Extreme NAS volumes can be created and deleted.
• The Config SysConfig parameter of EBS volumes is supported when you configure PVs.
• The issue that block volumes are loaded twice in BDF mode is fixed.
• EBS and NAS volumes are allowed to access APIs by using internal domain names.
• The Cloud Paralleled File System (CPFS) driver is upgraded and the dependency on the kernel is removed.

• EBS volumes can be unmounted before you delete the volumes.
• The disk creation policy is updated. Standard SSDs are created in preference to ultra disks. Ultra disks are created only when no standard SSD is available.
• UUID is supported as a high-priority search option to search for devices that use EBS volumes.
• The authentication management in managed Kubernetes clusters is updated.
• Security Token Service (STS) is supported to connect to OSS buckets.
• DuplicateMountPoint errors in EBS are fixed.
• The BDF protocol is supported to bind EBS volumes after the volumes are connected.

• Kubernetes clusters that use CSI and have no Internet access are supported.
• The issues related to mount path checks in EBS are fixed.

• The EBS PV name can be used as the disk ID. This feature is also supported by FlexVolume.
• Mount options can be configured for EBS volumes in MKFS Stage.
• Mount options can be configured to have a higher priority than the volume attributes of NAS volumes.
• Mount options of OSS volumes can be validated in OSS connectors.
• Subpaths of OSS buckets can be mounted as volumes.
• Volume topology can be used to dynamically configure LVM.