High Availability Overview

Velostrata Cloud Extension provides High Availability by using an active-passive model. Workloads use iSCSI multipath, and are connected to the primary and replica Cloud Edge nodes, but served only by the primary Cloud Edge node. When the primary Cloud Edge fails, the workloads failover to use the secondary Cloud Edge. Failback to the primary Cloud Edge is done once it becomes active again. For more information about the events that occur with failover and failback, see Viewing the Cloud Extension Status.

A Cloud Extension with an “impaired” status functions differently depending on whether one, or both, Cloud Edge nodes have failed:

  • If both Cloud Edge nodes are impaired, the Cloud Extension cannot be used to run new VMs in cloud, and will cause failures to VMs currently running in cloud. You can perform a forced move back, run the "prepare-to-detach" operation, and stop a VM.
  • If one of the Cloud Edge nodes is impaired, and the other Cloud Edge node is active most of the usual operations are still available.  

Similarly, if the VPN (access to CE), Cloud API, Backend or Management component fails, the Cloud Extension will be impaired and a limited set of actions will be available.

Use the tooltip to try and determine the reason for the failure. See Viewing the Cloud Extension Status.

Note: When one of the Cloud Edge nodes is down, the Cloud Edge does not provide high availability, in order to avoid any chance of data loss, all data is written to the object store, which reduces the Cloud Extension performance. For that reason, it is not recommended to move more VMs from on-premise to a Cloud Extension where one of the Cloud Edge nodes is impaired. When you move a VM to the cloud, and the Cloud Extension is impaired, the Cloud Extension name includes an indication that it is impaired in the Velostrata Cloud Extension page in the Run in Cloud wizard. See Viewing the Cloud Extension Status.

  • If a Cloud Extension is impaired due to incomplete deployment or initial health checks that did not pass. This is most likely due to pre-requisites that are not met, after fixing the underlying causes, use Repair Cloud Extension. Repairing the cloud extension attempts to re-create the missing components and/or run relevant health checks. Once the Cloud Extension is repaired, the status is set to Active.
  • If a Cloud Extension is impaired due to a cloud host health issue, stopping and starting the Cloud Extension may remediate the issue as affected instances are restarted on a healthy host. 

For more information on how to view the Cloud Extension status, and the events that occur, see Viewing the Cloud Extension Status.