Virtual Cloud Servers
We have created a cluster that meets all the requirements to be called cloud.
We use HPE MSA2060 and MSA2050 data storage systems with dual redundancy for all components. Shared external shared storage allows you to implement high availability for your virtual servers.
For SSD storage, enterprise-class SSDs are used, combined into many independent RAID10 + Hot-Spare.
HDD storage is based on HPE SAS drives, accelerated by an additional SSD layer (Performance Tiering).
Our clusters consists of high-end HPE DL360 Gen10 & Dell R640 servers with two 24-core Intel Xeon Gold 6428R CPUs
running at a high frequency of 3.0GHz, so even single-core virtual servers have sufficient performance.
The number of servers is kept at the N+1 fault tolerance level.
All cluster nodes, switches, and storage systems have 2 power supplies. Internet access is provided through 2 switches with different uplinks. Each of the two SAN fabric switches has 2 connections to each storage system. The latter, in turn, have 2 controllers operating in Active-Active mode. Each server, thanks to four separate network adapters, has a redundant connection to all switches, both to the Internet and to the SAN.
The software is as important as the hardware from which the cluster is built.
Our servers are clustered using software from industry leader VMware. The vSphere 8 software suite enables to fully utilize the power of the hypervisor server with 100% resource reservation. As a result, virtual servers run stably and even outperform entry-level physical servers in terms of power.
Thanks to the use of shared data storage in our cluster, such opportunities are available, like live migration of virtual servers. This allows real-time distribution of load on cluster nodes, therefore, there is no resources oversell!
Despite the fact that we use stable software on reliable servers, any hardware can fail. But in our cluster, this does not entail long downtime. Thanks to several independent algorithms of monitoring the health of the cluster, the system almost instantly sees the failure of any of the nodes and immediately reboots all virtual servers on other operating nodes.
Only true cloud servers can have such features!
Take a snapshot before invoking any action that could negatively affect the server operation. If something goes wrong, you will have the ability to instantly undo all changes with a few clicks and return the server to its original state. This feature is free!
You can access the server console directly from your browser. No plugins needed, you can even open the console from a mobile device. The web console will allow you to install OS or work with a server if the remote access services (SSH or RDP) fail.
Create a true failover cluster by linking the servers into a separate LAN! We have implemented very flexible settings - you can choose which servers should "see" each other. You can create up to 4 such networks, and by default all servers are in an isolated environment. Moreover, you can join servers from any of our locations into a single L2 network!
We make weekly backups of the entire virtual server free of charge. It's like a time machine - you choose the date of the restore point when you want to restore the data and click the corresponding button. Daily backups are also available for a nominal fee of $5/mo.