+31 (0)43 30 88 400 | office@comex.eu
COMEX X-ToolsSoftware by COMEX
Available

Shares and buckets that keep working

XHA is a gateway for SMB, NFS and S3 that writes every change synchronously to two volumes. If a storage system or an XHA node fails, your clients keep working, without losing acknowledged writes.

Clients (SMB / NFS / S3) Virtual IP XHA node A active XHA node B standby Witness synchronous Primary volume SMB · NFS · S3 Target volume SMB · NFS · S3

What XHA does for you

You pick a primary volume and a target volume. Clients connect only to XHA. XHA acknowledges a write only once both sides have it, and automatically switches to the other side when one fails.

Synchronous mirroring

Every write goes to the primary and the target volume. A client only receives an acknowledgement once the data is on both sides.

Automatic failover

Two XHA nodes in an active/passive cluster with a virtual IP address. A witness prevents split-brain.

Fast reconnect

After taking over the virtual IP, XHA signals known clients so they reconnect immediately instead of waiting for a timeout.

Resync and failback

A journal records what changed during an outage. Large files resume from the last checkpoint. Failback is automatic or manual per pair.

WORM and Object Lock

Supports WORM volumes and S3 with versioning, retention, legal hold and multipart uploads.

Protocol stays protocol

SMB to SMB, NFS to NFS, S3 to S3. No protocol translation, so the data stays readable on both systems.

Supported hardware

Built for FAST LTA Silent Bricks

We test X-Tools thoroughly on the hardware we know best. That is why we offer them primarily for FAST LTA Silent Bricks.

Discuss other hardware
  • FAST LTA Silent BrickTested and supported
  • FAST LTA Silent CubeSupport to follow
  • Other storage hardwareOn request, after a review and a test in your environment

XHA at a glance

Real screenshots of XHA, taken with demo data.

Two editions, one engine

XHA-Fast

The standard edition for FAST LTA Silent Bricks.

  • Primary and target are volumes on Silent Bricks
  • Tested and supported on Silent Bricks
  • Silent Cube support to follow

XHA

For other storage hardware, on request.

  • For SMB, NFS or S3 storage from other vendors
  • After a review of your environment and a test
  • Same features and engine as XHA-Fast

How to deploy XHA

  1. Two nodes and a witness

    Import two XHA appliances and a small witness VM. The cluster wizard joins them and sets up the virtual IP.

  2. Back ends and pairs

    Add your storage systems and create a pair of primary and target volume for each share or bucket.

  3. Switch your clients

    Point clients to the virtual IP of XHA. From then on everything is mirrored synchronously.

Technical details

ProtocolsSMB, NFS, S3 (always the same protocol for client and back end)
ClusterTwo nodes active/passive, witness as third vote, virtual IP
HypervisorProxmox VE; VMware and Hyper-V to follow
InterfaceWeb interface in Dutch, English and German
Trial30 days, all features, up to 2 pairs

Sizing guidance

LoadClient networkBack-end networkvCPU / RAMStaging disk
up to 1 TB/h (~280 MB/s)10 GbE10 GbE8 / 16 GB200 GB SSD
up to 4 TB/h (~1.1 GB/s)25 GbE2x 25 GbE16 / 32 GB500 GB NVMe
moremore XHA pairs per node pair

First guidance per node. Data is written to two back ends, so back-end bandwidth is doubled. We are happy to work out the sizing with you.

FAQ

Can I use XHA with storage other than Silent Bricks?

We offer XHA primarily for FAST LTA Silent Bricks, because that is the hardware we test and support it on. Silent Cubes will follow. For other storage hardware we can make arrangements on request, after a review of your environment and a test.

Does XHA use Silent Brick replication?

No. XHA works purely at protocol level (SMB, NFS, S3). This makes mirroring possible in both directions and XHA can technically work with storage from other vendors as well. We offer that variant on request.

What happens when a storage system fails?

XHA marks that side as unavailable and serves clients from the other side. Changes are recorded in a journal and synchronised after recovery.

What happens when an XHA node fails?

The other node takes over the virtual IP and the services. Thanks to the witness, two nodes can never write at the same time.

Can I test XHA first?

Yes. The trial runs for 30 days with all features and up to two pairs. Request it on the downloads page.

Ready to take the next step?

Want to know which tool fits your storage environment? We are happy to advise you, or you can download a trial right away.