Home

Cluster health · live performance · capacity overview
Not signed in
Standby

Cluster Overview Checking

Cluster Name
—
Health Status
—
Online Nodes / Engine
—
Total Disks
—

Storage Pool Capacity —

Pools
—
Total Capacity
—
Used
—
Topology
—

Recent Alerts

AI Auto Ops

Enable to auto-handle new and backlog alerts
Live Performance Monitoring Source: /api/live · adaptive sampling
Cluster IOPS
—
Target ≥ 100M · live
IO Throughput
—
Target ≥ 1 TB/s · live
Average Latency
—
Target < 0.01 ms · live
Storage Topology
—
—
Period

Live IOPS

Live IO Throughput

Latency Distribution Histogram

Node List

Online 0 Offline 0

Domestic Hardware Environment

  • Server—
  • CPU—
  • Disk—
  • Memory Hot Cache—
  • Node Scale—

Performance Targets (Live)

  • IOPS ≥ 100,000,000—
  • IO Throughput ≥ 1 TB/s—
  • Latency < 0.01 ms—
  • All nodes online—

Cluster Nodes —

Rank Map Native Ops

Rank is a storage engine instance. Shows membership state and SCM/NVMe health; Rebuild is listed separately as pool-level recovery.

Total
—
Joined
—
Degraded
—
Draining
—
Excluded
—
Evicted
—
Dead
—
Rebuild (pool-level recovery)
—

Rebuild is pool-level data recovery progress, not a Rank membership state (Joined / Degraded / Draining / …).

RankNodeHost / IPStatusSCMNVMeNUMAVersion
Loading…

Node Lifecycle Operations

RackNodeStatusMaintenanceEngineVersion
Loading…

Scale Out · Add Node Auto Provision

Enter SSH credentials for the new host; the system installs the storage node and joins the cluster. Expansion is gated by licensed capacity.

Rolling Upgrade

Current Version
—
Target Version
—
Status
—
Progress
—

Batch Configuration Deployment

Engine thread count maps to DAOS targets; SCM cache maps to ram scm_size. Illegal values are rejected. Changing targets/scm requires Allow reformat.

Export Cluster Configuration

Export current cluster configuration JSON for backup or cross-cluster migration.

Fine-grained Hardware Management Hardware Profile
Loading hardware data…
Loading topology…
ServiceStatusInstancesNotes
Loading…

Storage Pool Management

NameCapacityUsedUtilizationObject ClassFault DomainTieringQoSStatusActions
Loading…
POSIX data plane Same directory tree

Small updates stay on the fast tier. Large stable files move to the capacity tier, one file at a time, and an interrupted move resumes from the journal. A 3-replica or 2+2 pool still reads after one node or two disks are missing. NFS and S3 use that same POSIX directory.

No action yet
SASE adaptive placement Control policy

Media layout

Choose how data is placed. Saving changes the control policy only. It does not format disks or claim an HDD into a pool.

Hot-tier capacity

  • Loading…

Cold migration

Only one node may migrate at a time. Invalid watermark, benefit, or cost stays put.

Status
—
Slots left
—

Placement rules

  • Loading…
QoS & Tuning Managed Persist

Burst Performance Policy

Low-Latency Templates

Containers
NamePoolTypeObject ClassChecksumEncryptionEpochSizeUsedStatusActions
Loading…
Loading protocol status…
Performance Charts Shares live data with Home · switchable period
Period

Live IOPS

Live IO Throughput

Latency Distribution Histogram

Performance Targets

  • Cluster IOPS—
  • IO Throughput—
  • Average Latency—

Live sampling starts automatically when the page opens · source /api/live.

Microsecond-Level Deep Performance Analysis wall-clock
P50 Latency
—
P99 Latency
—
P999 Latency
—
P9999 Latency
—

Latency Percentile Curve (μs)

IO Path Latency Breakdown

  • Loading…

Filter Notes

Per-disk metrics support rack → Node → disk cascading filters. Without a node, show 1-disk overview per Node; with a node, show all 32 disks.

—

Per-Disk Performance

RackNodeDisk IDRead IOPSWrite IOPSRead BWWrite BWAvg Latency
Loading…

Snapshot List

NameContainerVOS EpochPersistentRollbackSizeTimeActions
Loading…

Scheduled Snapshot Policies P1

NameScheduleRetentionContainerStatus
Loading…

Data-level WORM Commercial Compliance

COMPLIANCE: no user (including admins) may modify/delete during retention; clearing legal hold requires expiry or dual control. GOVERNANCE: regular users cannot modify/delete; admins may override only after real password verification. Unified gate covers POSIX / S3 / Swift / FTP / WebHDFS / CIFS. Not hardware WORM / not a certified appliance; path-level immutability when true multi-version is unavailable.

Active
—
Compliance
—
Governance
—

Enable WORM Policy

Effective Query

IDScopeTargetModeRetain UntilLegal HoldStatusActions
Loading…

Clone Management P1

NameSource SnapshotSizeStatusActions
Loading…

Rebuild Control P1

Status
—
Progress
—
Bandwidth Limit
—

Remote Replication

SourceTargetModeStatusLagActions
Loading…

Multi-Tenant Management P1

NameQuotaUsedIOPS LimitBandwidth LimitStatusActions
Loading…

Local User Management P1

IDUsernameRoleStatusCreatedActions
Loading…

Roles P1

Loading…

LDAP / AD Integration P1

OIDC Configuration P3

Usage Reports & Billing P3

Total Capacity
—
Used
—
Utilization
—
30-Day Forecast
—

Capacity Trend

Period
TenantUsed TBQuota TBCost (CNY)
Loading…
TimeUserActionResourceResult
Loading…
Sovereign Compliance Dashboard Cross-Border Detection Push-button auditor validation · data boundary / telemetry ledger / sovereign key management

Data Boundary Ingress/Egress Network Monitor Data Boundary Monitor —

Home border: — · Baseline: — · VPN exempt: — · Violations: 0
Data Border
Local subnet nodes (10.x / 192.168.x) VPN tunnel (exempt) Foreign public IP (cross-border alert)
Active Conns
—
Local Subnet
—
Domestic public
—
Cross-border
—
VPN Exempt
—
In-border %
—
in-border vs cross-border connections (live)
—
Connection details (live)
NodeProtoLocal AddrRemote AddrDirStateBoundary
Loading…
Violation records
Last SeenNodeProtoRemoteCountry/RegionCountTag
No violation records

Air-Gapped Telemetry Log Verification Ledger Air-Gapped Telemetry Ledger —

  • Loading…
SeqTimeNodeLevelMessage (StorSoar engine checks / errors / kernel warnings)Chain HashExternal Sync Status
Loading…

Sovereign Key-Management (HSM) Status Ring Sovereign KMS / HSM —

— devices encrypted
—
  • Key Provider—
  • HSM/KMS Endpoint—
  • Master Key Local Path—
  • Certificate Pinning—
  • Third-Party Cloud Backdoor Key—
Audit CheckResultDetail
Loading…
Pooled NVMe Device Encryption States
DeviceNodeEncryptionKey Reference
Loading…
Anti-Ransomware Protection Center Behavior Detection · Auto Immutable Snapshots · Write-Budget Breaker · One-Click Recovery

Protection Status Protection Status — —

Loading…

Protection Overview Protection Overview

Immutable Snapshots
—
Anomaly Detections
—
Last Recovery Point
—
WORM Protected Dirs
—

Immutable Snapshot Vault Immutable Snapshot Vault

NameContainerCreatedWORM Lock ExpiryStatus
Loading…

Anomaly Detection Anomaly Detection —

Only a short burst counts as ransomware: mass extension changes, mass renames, high-frequency overwrites, a spray of ransom notes, or repeated snapshot deletes. Past the threshold, further changes are rejected, and an immutable snapshot is kept when auto-snapshot is on. Slow, small edits are left alone.

TimeTenant/ContainerSignalAction
Loading…

One-Click Recovery One-Click Recovery

Roll back the container to the immutable snapshot point. In native DAOS mode rollback may be unavailable (DER_NOSYS); the error is surfaced honestly.

Isolation & Air-Gap Isolation & Air-Gap

What storage can stop: mass encryption, a spray of ransom notes, and repeated snapshot deletes. Immutable snapshots cannot be quietly removed during retention, and WORM directories cannot be encrypted or deleted. The write-budget breaker returns 429 after 200 overwrites in 10 seconds.

What storage cannot stop: phishing mail, vulnerability exploits, RDP password guessing, a dormant backdoor, keylogging, or MBR tampering. Those happen on the host and never pass through the storage API. Files an account can already read, copied out slowly, are not treated as ransomware.

After an incident, disconnect, keep evidence, and report it. Paying the ransom does not guarantee a decryption key, and it does not guarantee the stolen copies are deleted.

Alert History

—
LevelMessageSourceTimeStatusAction
Loading…
—

One-Click Log Collection P1

Collect Engine/MDS/Monitor logs cluster-wide; real artifacts under /data/diagnostics.

Alert Notification Channels P1

Supports WeCom / DingTalk / Email / SMS, plus Slack, Microsoft Teams, Discord, Telegram, PagerDuty, and other common ops channels.

ChannelTypeStatusActions
Loading…

AI Auto Ops

Enable to auto-close new and backlog software-fixable alerts; physical items go to prioritized human todos
Event-driven auto-close + prioritized human todos (Critical/Secondary/Lower)
AI Ops Log
No logs yet
Enable option to auto-generate prioritized tasks

System Information

  • Product Version—
  • Data Persistence/data/mgmt/*.json
  • Live samplingAuto-starts when page opens
  • Mobile/mobile
API Reference

Full interactive docs are generated from FastAPI OpenAPI (parameters, bodies, auth). Sign in required; the list below is a curated summary.

    Automation Policy Engine P3 Scale-out · scheduled snapshots · self-healing

    Create Policy

    Loading…

    Compression & Deduplication P3

    • Loading…

    Recycle Bin P3

    NameTypeSizeDeleted AtActions
    Loading…

    K8s CSI P3

    OpenStack P3

    Prometheus P3

    
                

    Resource Tags P3

    TagBound ResourceActions
    Loading…

    Operations Config Templates P3

    NameTypeActions
    Loading…
    Security Settings presets + toggles China Default by default · switch to higher levels on demand

    Security Level Presets —

    Advanced Toggles

    Expand/Collapse toggles

    Effective Status

    —
    • Loading…

    Warnings

    • —
    License Management License 30-day trial (full features, unlimited capacity) · import RSA incremental capacity license after expiry

    Current License Status —

    Cluster ID
    —
    Licensed cap (TB)
    —
    Cumulative total capacity
    Used capacity (TB)
    —
    Usage
    —
    License type—
    Effective since—
    Trial remaining—
    Read allowed—
    Write allowed—

    Activate License

    Paste the vendor-issued license key below (requires admin login)

    Capacity in the key is the purchased increment; it accumulates with existing licenses. E.g. 100 TB + import +500 TB → 600 TB total cap.

    License Change History

    ActivatedKey digestDelta TBTotal TBOperatorResult
    Loading…
    —
    Loading node details…

    Disk Alert Handling

    0 total
    NodeSlotDisk IDHealthTempStatusReasonAction
    Loading…
    Switch rack/node to handle alerts on other nodes

    Handle