High Availability Strategies
What is High Availability?
By definition, High Availability (HA) is the ability of a system to operate continuously without failing for a designated period of time. HA works to ensure that a system meets an agreed level of operating performance. Data backup and fail-over capabilities play an important role in ensuring that HA systems meet their availability targets. MageMojo currently uses multiple backup strategies to ensure system stability and availability. In addition to back up concerns, system designers need to pay close attention to the data storage and access technology they use. Highly available systems must be well designed and thoroughly tested before they are used. The good news is, the experienced and knowledgeable MageMojo team has devised a system that will protect your data and grant you the peace of mind that accompanies knowing your site is stable.
Standard vs Private
|Database Replication||Real Time||Real Time|
|File Replication||4 hours||4 hours|
|Access to Read Only Replica||No||yes|
|Recovery Time||30 min to 3 hours||1 minute|
|Price||Free||Requires purchase of an additional stratus plan|
Standard High Availability
Standard High Availability maintains replication in a special replication container that runs using minimal memory as it only needs to write transactions that are already committed. Each of these containers is responsible for writing MySQL data for one stratus instance. This container is not accessible to the customer. Files are then replicated to the standby node every 4 hours. In the case of a catastrophic failure of an instance, the MageMojo team will first triage the failure and if appropriate, initiate the failover process that stops the appropriate containers for a failed store. The support team then redeploys the store where it mounts these backup filesystems. Because this component is manually done, there can be hours before all stores are failed over.
Private High Availability
Private High Availability utilizes a second Stratus instance which is available as a hot standby. This instance is accessible via the control panel and can also be used as a read-only instance for reporting or other read operations. Files are also replicated to this instance every hour. In the case of a catastrophic failure of an instance, there is a watcher that monitors the status of all Private High Availability stores. If a store goes into a “Not Ready” state, this watcher will initiate a failover of that store and switch the primary and failover instances. After this switch, the watcher will then repoint CloudFront to standby, bringing the store back online within one minute.