High Availability

High availability (HA) provides automatic failover between two nodes using a virtual IP (VIP). Marketplace deployments configure HA automatically; this page covers manual HA setup (on-prem) and ongoing management.

Access

Navigate to Failover in the Web UI sidebar.

HA Modes

Mode	Description
Active-Passive	One resource group owned by the active primary node. Resources transfer only during failure or manual intervention.
Active-Active	Two resource groups with different cluster IDs. Each node owns one resource group. On failure, the surviving node assumes both.

Storage Models

HA works with either shared or shared-nothing storage:

Model	Description
Shared Storage	Both nodes access the same storage (dual-path arrays, NVMe-oF, cloud block storage)
Shared-Nothing	Separate storage per node with synchronous mirroring (NVMe-oF mirror or DRBD)

Prerequisites

Cloud-Specific Requirements

Cloud	Requirements
Azure	Service principal login on all nodes, Availability Set, Load Balancer or custom route for VIP
GCP	Service Account with custom IAM roles, secondary IP ranges configured
AWS	Custom IAM role with EC2 permissions, subnet with secondary private IP support
On-Prem	Network allowing gratuitous ARP for VIP, shared storage or dedicated replication network

General Requirements

• Unique cluster ID (1–255) for all managed resources
• Node names must match uname -n output
• A virtual IP address (outside normal DHCP range)
• At least one heartbeat link between nodes (two recommended)
• Clear previous disk headers with wipefs if reusing disks

Configure Failover

Set Up via Web UI

1. Navigate to Failover and click New
2. Complete the cloud-specific checklist (if applicable)
3. Configure:

Field	Description
Failover Type	Active/Passive or Active/Active
Primary Node	Primary hostname (auto-populated)
Secondary Node	Secondary hostname (must match uname -n)
Resource ID	Cluster resource ID (1–255)
Virtual IP	Floating IP address for client access
Netmask	Subnet mask (on-prem only; cloud auto-configures)

4. Configure heartbeat links:

Field	Description
Heartbeat Link 1	Primary network interface + peer IP
Heartbeat Link 2	Secondary interface + peer IP (recommended)
Ping Node	Gateway IP for split-brain detection

5. Optionally configure advanced parameters:

Parameter	Description	Default
Keepalive	Heartbeat interval (ms)	1000
Deadtime	Time before declaring peer dead (s)	10
Warntime	Warning threshold (s)	5
Initdead	Initial dead time for startup (s)	30
Auto Failback	Automatically failback when primary recovers	Off

6. Click Create (takes ~30 seconds)

Join an Existing Cluster

If the primary is already configured, use Join on the secondary node to add it to the cluster.

Manage Failover

Action	Description
Start	Start the failover service on a node
Stop	Stop the failover service
Standby	Release all resources (go standby)
Takeover	Force the other node to release all resources
Delete	Remove the failover configuration

Monitor Status

The Failover panel shows:

• Node list with roles (Primary/Secondary)
• Failover service status (running/stopped)
• Virtual IP assignment
• Resource ownership per node
• Heartbeat link health
• Per-node active/inactive status

CLI

# Create failover configuration
mayacli create failover \
  node=server1.local,server2.local \
  options="clusterid=3 ipaddress=10.1.1.111 \
  hblink=eth0>10.1.1.101:eth0>10.1.1.100,eth1>192.168.100.101:eth1>192.168.100.100"

# Show failover status
mayacli show failover

# Start failover service
mayacli bind failover

# Stop failover service
mayacli unbind failover

# Go standby (release all resources)
mayacli failover standby

# Force takeover (take all resources)
mayacli failover takeover

# Takeover specific: local, all, foreign, or failback
mayacli failover takeover local
mayacli failover takeover failback

# Delete failover configuration
mayacli delete failover

See CLI Guide for the full command reference.