This summer, we launched #CloudMTS backup product based on Acronis solution in test mode. It took us less than two working days: installation, commutation and update of the equipment required. Thanks to the experience of our specialists and selected equipment, we managed to launch the project quickly and easily: excellent guides and convenient ARM kits. If you assembled everything correctly, then even one engineer is able to reconnect the shelf without problems, replace the disk and, in general, perform any work that is not related to dismantling the shelf.
In this post we will discuss what physical base lies at the core of Acronis-based #CloudMTS storage. But first, let’s talk about why customers need backup services, and what are the advantages of such Acronis-based solutions.
Client’s use cases
Case #1. The client gets the option to use and access a highly reliable and always readily available service for creating, storing and managing backups, as well as restoring them, if necessary, in the #CloudMTS cloud main backup system.
There is no need to spend time and resources on purchasing equipment for storing backup copies, which, moreover, with a high degree of probability will not be completely loaded and, therefore, will stand idle. A client can simply order the required resources for solving current issues. If necessary, the volume of allocated equipment can be changed. Access to the Self-service panel is granted to manage backup. The panel has a predefined Russian language and a user-friendly interface, allowing to configure the backup process in just three clicks (Figure 1).
Case #2. An additional feature is free migration from the client’s stage to #CloudMTS cloud. The advantage of using Acronis in this scenario is the ability to transfer data regardless of the source hardware or virtualization platform. For example, a typical case is migration from Hyper-V to VMware in Elastic Cloud (IaaS). Using Acronis-based backup from #CloudMTS, this process does not take much time and effort – you need to backup a virtual machine of one hypervisor (for example, Hyper-V), and then select this backup for recovery on the Elastic Cloud (IaaS) host. The process is fully automated if we talk about VMware and Hyper-V. A virtual machine is created and disks of the right size are added. In case when it is needed to transfer data from physical servers, the same scenario as described above is applied, but instead of the virtual machine, the original machine is a physical one.
Case #3. Automation of machines mass configuration is another scenario when #CloudMTS can be used. With Backup-as-a-Service it is possible to significantly reduce the time that administrators spend on configuration of each machine – all settings are carried out centrally from a single repository based on the configured policies.
Case #4. Backup to #CloudMTS cloud will not only allow to create, store, manage and restore backups, but also protect from encryption programs and crypto miners. This solution is based on Acronis Active Protection technology.
Let’s review an example of how Acronis Active Protection technology works
The user receives an e-mail from the attacker, containing a link, through which an inattentive user or an employee, infects his computer with ransomware (software that encrypts data, after which it requires a ransom, usually an online payment for data decryption). If Acronis Agent is installed and Acronis Active Protection is activated (this function is disabled by default) on the machine then all suspicious activity will be detected with the help of behavioral analysis. Files that the virus is trying to modify will be instantly saved and all modification processes will be blocked. Even if some file was damaged, as a result of the attack, it will be restored from the backup automatically.
As a result, the total installation throughput is a little less than 30 GB/s, and when using the protection against failure of 2 physical nodes, it is 23 GB/s.
What are the specs?
Acronis solution from #CloudMTS is deployed on the basis of Huawei 2288V5 and WD HGST Ultrastar Data60. The basic installation contains 6 Huawei 2288V5 servers, each has an Avago MegaRAID SAS 9380-8e + CacheVault LSICVM02 connected to PCI-Express 3.0 x8, through these controllers WD HGST Ultrastar Data60 1ES0352 SE4U60 are connected. One server equals one shelf, respectively.
What is WD HGST Ultrastar Data60 and why does #CloudMTS use this particular solution?
Ultrastar Data60 is an enterprise-class hybrid shelf. Suitable for work with large amounts of data, especially with write-intensive patterns. They can be used in data centers where there is a need for high storage density, high availability and low power consumption.
This solution is cheaper and faster than a data-storage system.
The new disk array allows to build a storage with several petabytes and dozens of GB/s of performance way cheaper than using horizontally scalable systems. Unique WB technology allows to reduce vibration by 60% compared to competitors, which allows to extend the life of HD. This is a new technological breakthrough in cooling. It creates air ducts inside the chassis that allow cold air to be drawn to the rear rows of disks directly from the cold corridor, bypassing the front rows. As a result, the fans run much quieter, the disks live longer and work faster.
This solution is suitable for following tasks:
The solution is promoted as a hybrid storage, with a 4U form factor, the maximum possible capacity at the moment is 840 TB, it can be equipped with 24 SSD disks (SAS 12 GB/s or SATA 6 GB/s) to create an acceleration pool. There are two IOM modules inside operating in a fault-tolerant configuration; 6 mini-SAS HD ports are accessed externally from each IOM module. Technical specifications available here.
WD HGST Ultrastar Data60.
WD HGST Ultrastar Data60.
The photo represents the installation in one of #CloudMTS data centers.
We will keep all installation layers a secret, but each server is a node of Acronis Cyber Infrastructure solution (formerly Acronis Software-Defined Infrastructure, which we looked at 1 year ago), disk storages are connected to the servers through SAS MegaRAID 9380-8e controllers, controller is in JBOD mode, and accordingly the operating system sees each of the 60 disks, no RAIDs are used.
We have assembled and successfully use the following architecture
Installation turned out to be very productive. The resulting bandwidth of each physical node is 6.3 GB/s, here we rest on the SAS connection bus of the PCI-Express 3.0 x8 controller with a laboratory bandwidth of 7.88 GB/s. The symbiosis of Huawei and WD HGST Ultrastar Data60 servers with 60 disks of 9 TB each allow us to get 2 PB of active capacity with excellent performance, protection from failure of 2 physical nodes and a competitive cost for 1 GB storage.
Product Details here
Text prepared by: Dmitry Sudarikov, lead architect, Sergey Sklabovsky, product manager, Anton Kalinkin, head of Professional and Managed Services, and Sergey Shashin, product manager.