Cluster & High Availability
Multi-node Proxmox clusters with Corosync quorum, HA groups and automatic failover, so a single server failure is a non-event.
Included in this service
Three-node minimum, properly
Quorum needs an odd number. Two nodes and a witness is a compromise we will explain rather than hide.
Corosync on a dedicated link
Cluster traffic gets its own network so a busy backup cannot trigger a false failover.
HA groups by priority
Critical VMs restart first, on the node you nominated, in the order you decided.
Fencing configured and tested
A node that stops responding is isolated, so two copies of the same VM never write to one disk.
What changes for you
Maintenance during working hours
Move workloads off a host, patch it, move them back — nobody notices.
A host failure is an email, not an outage
Workloads restart elsewhere automatically, usually inside two minutes.
Tested, not assumed
We pull a power cable during handover so you have seen it work.
From first call to handover
-
Assess
We audit what you run today — hardware, licences, workloads, pain points and growth plans.
-
Design
A written design with sizing, redundancy, network layout and a clear cost breakdown.
-
Implement
Build and configure in a staging window, with nothing pointed at production until it is proven.
-
Migrate
Workloads move in planned batches, each one verified before the next begins.
-
Handover
Documentation, credentials, runbooks and training for your team.
-
Support
Ongoing monitoring, patching and support under an agreed response time.
Asked often enough to answer here
Do we need shared storage for HA?
For automatic failover, yes — Ceph or a shared array. Without it you can still do live migration between hosts, but a failed host means restoring from backup rather than restarting elsewhere.
How long does failover actually take?
Typically 60–120 seconds: the cluster has to be certain the failed node is gone before it starts the VM elsewhere. Being certain is what stops data corruption.
Talk to an engineer about cluster & high availability
Tell us what you run now. You get a written design and a fixed scope before any invoice.
Also in IT Infrastructure
Proxmox VE Virtualization
Consolidate ageing physical servers into a managed Proxmox VE environment — lower hardware cost, no per-socket licensing.
Storage Replication & Backup
ZFS and Ceph replication with scheduled, verified backups through Proxmox Backup Server and off-site copies.
Disaster Recovery Management
Defined RTO and RPO, a secondary site, tested restore drills and a documented runbook your team can actually follow.