Recommendations on how to recover data from VMware safely

• Extract backups to another partition. It will help to make sure that all backed up files are operable before possible rewriting of data on a functioning partition.
• If RAID array failed, test backup copy on another drive or create an image of each RAID disk before starting array rebuild. Sometimes RAID rebuild works incorrectly and it may exacerbate the problem.
• Do not create new files on the disk from which you intend to recover data as well as discontinue using virtual machines until the important data is recovered. New files may rewrite the files you need to recover if extraction from backup copy is not successful. Running virtual machines that use snapshots and thin virtual disks also may rewrite files that you need to recover.

In this article I will tell you about one of the features of Veeam Backup & Replication 6.5, namely that which provides effective VM (virtual machine) recovery from SAN snapshots.

HP LeftHand Storage

Virtual environment provides opportunity to run several VMs on a single server. Usually when a company expands, amount of its servers and VMs increases too. As a result capacity and performance of SAN does not longer suffice. A common solution is to add disk enclosures until a pair of controller units becomes a weak spot in the infrastructure development. If such problem occurs, a company upgrades a pair of controller units or installs a new SAN.

From the point of view of its architecture, HP SAN Left Hand Storage (HP SAN) differs considerably from traditional SANs in capacity and performance. HP SAN has no separate controllers or disk enclosures. However, each array has its own resources (processor, HDDs, cache memory, I/O ports). That’s why besides storage capacity, addition of new arrays facilitates higher processor performance, provides additional cache memory, etc.

So, HP SAN is a cluster of storage nods. Each nod is a fully-featured data warehouse. Data are distributed automatically between nods. This process is controlled by a single console. Addition of new nods facilitates increase of performance and reliability of HP SAN.

The article reviews types of virtual hard disk drives in Microsoft Hyper-V.

Virtual Hard Disk Controllers

Windows accesses hard disk drives via controllers and Hyper-V-based virtual machine is not an exception.

In virtual machines on can choose either IDE or SCSI controllers.

IDE Controller. Hyper-V uses emulated devices with IDE Controller. One can install two IDE Controllers with two disks in each controller. Boot Drive should be connected to one of the IDE devices. Boot Drive can be either a Virtual Hard Disk or a physical disk. Although in a virtual machine in order to start a guest operating system boot drive should be an IDE device, there are many types of physical devices among which one can find a storage for IDE device. For example, one can use one of the types of physical storage devices listed at the beginning of this section.

SCSI Controller. Each virtual machine supports up to 256 SCSI disks (four SCSI controllers, each supporting up to 64 disks). SCSI Controller uses a type of devices that was specially developed for virtual machines and that uses virtual machine bus to exchange data. Virtual machine Bus is available after the start of Guest Operating System. Therefore virtual hard disks, connected to SCSI controllers, cannot be used as Boot Drives.

The technology that is used for SSDs is not just used for hard drives. Other popular storage devices, including USB thumb drives and camera (or phone) memory cards use the same technology as SSD drives. New technology is always a great thing, but it can also pose problems, some of which the average user isn’t even aware of.

How Solid-State Drives Delete Data

Solid-State Drives don’t work anything like the traditional magnetic platter drives we’re used to. Those platter drives can read, write, and erase anywhere on the disk at any give time. Data is regularly overwritten as new data needs room. This is not the case with SSDs.

SSDs use NAND flash technology to save data. NAND dictates that individual cells be aligned in rows of cells called pages. This is the smallest amount of data that can read or written. A page is approximately the same size as a cluster on a magnetic platter hard drive. A group of pages make up a block, and the block is the smallest unit that can be erased on an SSD. When an SSD needs to write new data, it can only write data to a ‘blank’ page. NAND technology does not allow overwriting of data the way that magnetic platter hard drives do. Since you can write in a page, but can only erase in a block, if the entire page is not blank, and the SSD needs that page, any data on the entire block that needs to be kept has to be moved to a different block and the entire block erased so that the new data can be written.

Increasing performance of disk subsystem is the most pressing issue of today. The reason for this is that HDD, considering its low price, still occupy the leading positions in mass segment. Spindle disks become the bottleneck in more than half of resource-intensive applications. In this case everything depends not on SATA throughput interface, but on the physical capabilities of mechanical parts of magnetic disk. Throughput of SATA-II and SATA-III interfaces is 300 MByte/s and 600 MByte/s respectively. Maximum performance that can be provided by an ordinary HDD does not exceed 150 Mbyte/s. That is why shifting to SATA-III interface is reasonable only for SSD, through not for all.

To assess the efficacy of disk subsystems of different types, we selected the following solutions for testing:

1. OCZ RevoDrive X2 PCI-E SSD 100Gb

Many owners of solid-state drives have a reasonable question "Why declared in specifications read/write speed of solid-state drive is less than what they have in reality?" It could be caused by a number of reasons described below. But it has to be mentioned that usually maximum speed is mentioned in SSD specifications.

In this article you will find information about controllers used in manufacturing solid-state drives and list of solid-state drives based on the particular controller (and memory chips). There is a controller in every SSD. Controller contrains the electronics that bridge the NAND memory components to the host computer. The controller is an embedded processor that executes firmware-level code and is one of the most important factors of SSD performance. The controller performs the following main functions:

• Error correction (ECC)
• Wear leveling
• Bad block mapping
• Read scrubbing and read disturb management
• Read and write caching
• Garbage collection
• Encryption

The list of controllers is as follows: