Skip to main content

Command Palette

Search for a command to run...

Hwameistor is an HA local storage system for cloud-native stateful workloads.

Updated
7 min read

HwameiStor

HwameiStor is an HA local storage system for cloud-native stateful workloads. It creates a local storage resource pool for centrally managing all disks such as HDD, SSD, and NVMe. It uses the CSI architecture to provide distributed services with local volumes and provides data persistence capabilities for stateful cloud-native workloads or components.

System architecture

Current Status

CNCF logo

HwameiStor is a Cloud Native Computing Foundation sandbox project.

The latest release of HwameiStor is

Release Status

See current releases.

Running Environments

Kubernetes compatibility

kubernetesv0.4.3\>=v0.5.0\>= 0.13.0
\>=1.18&&<=1.20YesNoNo
1.21YesYesNo
1.22YesYesNo
1.23YesYesNo
1.24YesYesYes
1.25NoYesYes
1.26NoYesYes
1.27NoNoYes
1.28NoNoYes
1.29NoNoYes
1.30NoNoYes

Modules and Code

HwameiStor contains several modules:

local-disk-manager

local-disk-manager (LDM) is designed to hold the management of disks on nodes. Other modules such as local-storage can take advantage of the disk management feature provided by LDM. Learn more

local-storage

local-storage (LS) provides a cloud-native local storage system. It aims to provision high-performance persistent LVM volume with local access to applications. Learn more

Scheduler

Scheduler is to automatically schedule a pod to a correct node which has the associated HwameiStor volumes. Learn more

admission-controller

admission-controller is a webhook that can automatically determine which pod uses the HwameiStor volume and, help to modify the schedulerName to hwameistor-scheduler. Learn more

Evictor

Evictor is used to automatically migrate HwameiStor volumes in case of node or pod eviction. When a node or pod is evicted as either Planned or Unplanned, the associated HwameiStor volumes, which have a replica on the node, will be detected and migrated out this node automatically. Learn more

HA module installer

DRBD (Distributed Replicated Block Device) is one of third-party HA modules which the HwameiStor will leverage to provide HA volume. It composed of Linux kernel modules and related scripts to build high available clusters. It is implemented by mirroring the entire device over the network, which can be thought of as a kind of network RAID. This installer can directly install DRBD to a container cluster.

Exporter

Exporter will collect the system metrics including nodes, storage pools, volumes, disks. It supports Prometheus. Learn more

Volume Snapshot

HwameiStor provides the feature of snapshot and restore on the LVM volumes. Currently, the snapshot/restore feature works for LVM non-HA volume. Learn more

Volume Auto Resize

HwameiStor can automatically expand the LVM volume according the pre-defined resize policy. User can define the preferred policy and describe how and when to expand the volume, and HwameiStor will take the policy into effect. Learn more

Volume IO Throtting

HwameiStor can set a maxmium rate (e.g. bandwidth, IOPS) to access a volume. This feature is very important to prevent the Pod from crashing, especially in the low-resource condition. Learn more

App Failover

The feature of failover is to actively help the application to fail over to another health node with the volume replica, and continue the working. Learn more

Audit

HwameiStor provides the information about the resource history, including cluster, node, storage pool, volume, etc. Learn more

UI

HwameiStor provides a friendly UI to the user to operate the cluster. Learn more

Documentation

For full documentation, please see our website hwameistor.io.

For detailed adopters that have HwameiStor deployed in a production environment or a user acceptance testing environment, please check the adopters list.

Build and Preview Website Locally

It's a good practice to build and preview your modifications locally before submit a PR to HwameiStor Website. You can run the following commands to preview locally:

cd docs
npm run start  # for english
npm run start -- --locale cn # for i18n/cn

Change Navigation

To change the navigation (left sidebar) sequence, you can change the content in _category_.json that exists in each folder:

{
  "label": "Installation", // The name displayed on the left sidebar
  "position": 4, // The sequence identifier, 1 is listed at top
  "link": {
    "type": "generated-index",
    "description": "In this section, we will introduce the installation procedure:"
  }
}

For i18n/cn, you can change the nav in current.json:

{
  "sidebar.tutorialSidebar.category.Modules": {
    "message": "Modules", // The section titles displayed on the left sidebar
    "description": "The label for category Modules in sidebar tutorialSidebar"
  },
  "sidebar.tutorialSidebar.category.Modules.link.generated-index.description": {
    "message": "This chapter introduces the following modules included in HwameiStor:", // The description before sub-section cards
    "description": "The generated-index page description for category Modules in sidebar tutorialSidebar"
  }
}

Roadmap

FeaturesStatusReleaseDescription
CSI for LVM volumeCompletedv0.3.2Provision volume with lvm
CSI for disk volumeCompletedv0.3.2Provision volume with disk
HA LVM VolumeCompletedv0.3.2Volume with HA
LVM Volume expansionCompletedv0.3.2Expand LVM volume capacity online
LVM Volume conversionCompletedv0.3.2Convert a non-HA LVM volume to the HA
LVM Volume migrationCompletedv0.4.0Migrate a LVM volume replica to a different node
Volume GroupCompletedv0.3.2Support volume group allocation
Disk health checkCompletedv0.7.0Disk fault prediction, status reporting
LVM HA Volume RecoveryPlannedRecover the LVM HA volume in problem
HwameiStor OperatorCompletedv0.9.0Operator for HwameiStor install, maintain, etc.
ObservabilityCompletedv0.9.2Observability, such as metrics, logs, etc.
FailoverCompletedv0.12.0Fail over the pod with HwameiStor volume
IO throttlingCompletedv0.11.0Limit IO bandwidth to access the HwameiStor volume
Disk replacementPlannedReplace disk which fails or will fail soon
LVM volume auto-expansionCompletedv0.12.0Expand LVM volume automatically
LVM volume snapshotCompletedv0.12.0Snapshot of LVM volume
LVM volume cloneCompletedv0.13.1Clone LVM volume
LVM volume thin provisionUnplanedLVM volume thin provision
LVM volume stripe modeUnplanedLVM volume stripe read/write
Data encryptionUnplanedData encryption
System ConsistencyPlannedConsistent check and recovery from a disaster
Volume backupPlannedBackup the volume data to remote server and restore
HwameiStor CLI commandCompletedv0.12.4CLI command is to manage the HwameiStor cluster
HwameiStor GUICompletedv0.11.0Manage the HwameiStor cluster
Cache VolumeCompletedv0.14.5Accelerate dataset loading on node

Community

We welcome contributions of any kind. If you have any questions about contributing, please consult the contributing documentation.

Blog

Please follow our weekly blogs.

Slack

If you want to join the hwameistor channel on CNCF slack, please get invite to CNCF slack and then join the #hwameistor channel.

WeChat

HwameiStor tech-talk group:

QR code for Wechat

Discussion

Welcome to follow our roadmap discussions here

Pull Requests and Issues

Please feel free to raise requests on chats or by a PR.

We will try our best to respond to every issue reported on community channels, but the issues reported here on this repo will be addressed first.

More from this blog

SRE Tribe

413 posts