Hello, friends! In the last article we discussed what to do if one hard drive is already filled with files and we need to create a mirror for it. I propose to do this today. Before starting, I will briefly remind you of what a RAID array or mirroring is.

The principle of operation of a RAID array is duplication of information; in simple words, your computer will use two hard drives to store files, which will completely copy each other; if you wrote a file to the first hard drive, it will also be copied to the second drive. This is done for the security of your information and if one hard drive suddenly breaks down, then all files will remain safe and sound on the other hard drive! The only drawback of a RAID 1 array is that your two hard drives will work as one, for example, when installing two 1TB hard drives into the system unit, in the operating system they will both be defined as one 1TB hard drive.

• Note: Read the next article and this section ""

So, let’s imagine a situation: you have two hard drives installed on your computer: a solid-state drive with Windows 8.1, and also a simple 250 GB hard drive with the most important files that you should never lose, so let’s create the simplest RAID 1 is an array of two hard drives, that is, we buy another 250 GB hard drive and install it in the system unit.

After that, turn on the computer and after loading the operating system, go to “Disk Management” and see three hard drives:

Disk 0- SSD solid state drive, drive C: with Windows 8.1.

Disc 1- a regular HDD (New volume (D:) with a capacity of 250 GB, with your files, for which we will create a mirror.

Disc 2- clean HDD, also with a capacity of 250 GB, will be a mirror of Disk 1.

The volume of the disks does not have to be the same, the main thing is that the mirror is not smaller in volume than the disk from which it is created.

Right-click on Disk 1 and select Convert to Dynamic Disk.

Make sure the disk is selected correctly. OK.

Convert

Disk 1 (New Volume (D:) has been converted to a Dynamic Disk, nothing happened to our files, they are available.

Click on New volume (D:) right mouse and select Add a mirror,

Select Disk 2 with the left mouse and click on the Add mirrored volume button.

The process of synchronizing the contents of hard drives is taking place, all information from The new volume (D:) is copied to the mirror.

"Disk Management" reports that synchronization is complete, the disks are healthy and you can work.

This PC window RAID 1 array is presented as one volume.

It is customary to write a success story on corporate blogs - this has a positive effect on the company’s image. Unfortunately, everything in an engineer’s work does not always end with a happy ending.
I must say that my colleagues are already starting to joke that I “attract” problems. In one way or another, I have participated in almost all problematic applications recently. And now I want to tell you one instructive story from my practice.

The story began with the fact that I was asked to analyze the performance of a disk array of one storage system, the “brakes” of which paralyzed the work of an entire branch. The starting situation is this:

• The array contains VMware farm datastores.
• All volumes are located on RAID5 (7200 and 10000 disks) and are mirrored between two identical arrays.
• There is no contract with Vedor to support this equipment.
• The array firmware version is 7.3.0.4 (current at that time 7.6.1.1).
• The storage system is also used for virtualization of HP EVA storage systems.

According to the array performance logs, the “brakes” did not occur due to increased load. I suspected that the cause of the problems was a failed controller on the HP EVA virtualized storage system. Usually performance problems are solved remotely, but in this case they decided to send an engineer to the site (at that time no one suspected that the business trip would last two weeks).

And then, during the performance analysis, a “poltergeist” began to appear: volumes from the array in the vSphere interface periodically display an incorrect volume (from negative to tens of petabytes), which the customer regarded as a problem in the array. At the same time, access to the consoles of some virtual machines was lost, and other troubles arose. Even I was already starting to get nervous, and the customer was simply out of touch.

And then the fireworks of problems begin.

We found a bug in ESXi that may cause incorrect volume sizes to be displayed. But it turns out that there is no official contract for VMware support. Support is provided by a third-party company and only on weekdays, and it happens on Saturday.

For complete happiness, the firmware of two of the three servers and switches (blade chassis) lags behind the firmware of the chassis control module, which can also lead to the most unexpected problems. Well, the cherry on the cake: SAN switches have different firmware versions, and all of them are of the major version before last (6.x.x, when 8.0.x is available).

Finally, it turns out that MS SQL Server Express ran out of free space, which caused a “poltergeist” with the availability of VM consoles in vSphere and volume sizes were displayed incorrectly. So while the administrators were solving database problems, we were trying to figure out the storage system.

After some actions, the main volume suddenly went offline.

We remembered a bug in storage system firmware versions 7.3, 7.4 and 7.5, due to which bad blocks may appear on compressed volumes after a certain number of accesses (in this situation, neither RAID fault tolerance nor mirroring volumes to a neighboring array can help, since the error located at a higher level).

And here the most interesting nuance appeared: it turns out that the customer’s IBS has not been working for 3 months. That is, there are backups, but they are not current, and restoring from them is the same as losing data.

We managed to bring the volume online (via the array CLI), but the first time the host tried to write something, it crashed again. We turned off all the datastores on the servers and spent the next day in the office, almost without breathing, copying all the virtual machines wherever possible - to servers, USB drives and PCs.

As a result, we managed to save all the data except the VM on which we launched snapshot consolidation, since during the consolidation process the LUN went offline, and instead of the VM data, we were left with “porridge.” According to the law of meanness, it turned out to be an electronic document management VM. In addition, to eliminate various risks, we had to update almost the entire infrastructure - VMware, Brocade, HP Blade, and so on.

Prerequisites for the disaster

What conclusions can a dear reader draw from this story so as not to find himself in a similar situation?

Thank you for your attention, work smoothly for you.

Alexey Trifonov

Backing up information is one of the main ways to protect it from loss due to equipment failure.

A good place to start is to differentiate between backups systems and backup separate files.

Here are a few simple rules that will help you save information.

1. Copy frequency.

Let's look at operating system backup first.

Creating a backup copy will help you avoid installing the entire system, configuring it, installing programs, and so on. With the improvement of computer technology, these methods become irrelevant, because any operating system itself creates backup copies and restores them in case of failure.

Backup other information.

For those who constantly subject their system to any changes that could “kill” it, it is best to make backup files every time before starting such work. It will be quite a shame to lose the fruits of your work in a few days, forgetting to make another backup before the system crash. Of course, it is not necessary to make a backup before each installation of any new program. Typically, it is recommended to make a backup if you need to save any new data, the creation of which took a lot of time and money. For ordinary users, one backup per month is enough.

2. Storage.

Before making a backup, the question arises - where to store the information?

First you need to understand how important and confidential this information is. If the information is of value only to the user himself, then the following methods can be used.

a) Of course, you can store a backup copy on the computer itself if the hard drive is divided into several logical ones. In this case, the backup copy is saved on any disk except the system one. It is the system disk on which the operating system is located that is most often susceptible to various failures from programs installed on it.

b) Storing on external media, for example, on a flash drive or on a removable disk, will protect you from loss of information if the entire hard drive in the computer fails.

c) Storing information on the Internet.

Network drives, such as Yandex drive and others, are becoming increasingly popular. Data storage is quite convenient, but not secure.

If the information contained in the backup is of interest not only to the user, then you should think about the security of access to it. Any external media can simply be stolen.

3. Data verification.

After a backup is made, you need to check whether the information is in the backup copy and whether it can be used. Simply put, the user may make a mistake and back up the wrong file.

Built-in OS capabilities.

as mentioned above, any operating system itself creates backup copies of system files.

Consider this function using the Windows 7 operating system as an example. It is called “Backup and Restore Center”.

To launch the Backup and Restore Center utility, do the following:

In the window that opens, you can make a complete copy of the operating system, make a copy of individual databases, or restore files after a system failure.

This service fully satisfies the needs of ordinary users.

After a failure, all you need to do is find your backup copy, run it, and then the system itself will tell you what actions to take.

If the user is still not satisfied with the standard method of creating backups, then there are many paid and free utilities that will help make a backup copy.

Here are the most popular programs for creating backup copies.

Top three paid programs:

1. Norton Ghost
2. Paragon Backup & Recovery
3. Acronis True Image Home

Free programs:

• FBackup 4.8
• File Backup Watcher Free 2.8
• Back2zip 125
• The Copier 7.1
• Comodo BackUp 1.0.2

From the list of free programs listed, you can choose any one that is most suitable for the type and activity of using the computer. The weakest copying program is File Backup Watcher Free 2.8, but this program has one big advantage - creating ISO images. Back2zip is suitable for those who rarely have to deal with backups and essentially have almost nothing to copy. The Copier is quite complicated to use, but can help you reserve up to 300 gigs of data.

Comodo BackUp is one of the professional programs that will help you create backup copies of documents, configure settings for automatically creating backup copies and sending them to an external resource or FTP server.

Disk cloning.

Probably everyone has encountered the problem when the disk runs out of free space.
What to do if there is nothing left to delete, but there is still not enough space?
You have to buy a new hard drive with a larger capacity. It’s good if there is room in the system unit for a second hard drive, but what if you can only replace one hard drive with another? You need to somehow transfer all the data from the old hard drive to the new hard drive. Special disk cloning programs allow you to do this.

The most common one is Acronis 2011, which helps to perform high-quality disk cloning.

This program has 2 cloning modes. Manual and automatic.
In manual mode, the user can select the areas that need to be copied. The progress of the process is displayed in the program window. After completing the program, you will be asked to restart the computer, after which you can replace the old hard drive with a new one.

The next most popular program is HDClone. The principle of operation is almost the same. The only difference is the price of the product and slightly reduced functionality.

The third place is taken R-Drive Image

A fairly easy to use program with a step-by-step user interface. The main advantage is the low cost of this product.

There are also many free similar programs that are not inferior in functionality to the above programs. An example of such a program would be Clonezilla and PC Disk Clone Free 8.0. There is also a paid analogue of the second program.

RAIDarrays. No, not insect repellent.

RAID was created in 1987 by A. Petterson, A. Gibson and Katz. Initially, RAID - “redundant array of inexpensive disks” was translated as “a spare array of inexpensive disks.” Later, with the increase in the price of hard drives, RAID began to have the meaning of "redundant array of independent disks", i.e. "spare array of independent disks."

Previously, RAID arrays were used only for servers, but now with the constant development of technology, RAID arrays are also used for home computers.

A RAID array is designed to speed up computer operation and increase the reliability of data protection and storage. Depending on the configuration of the RAID array you choose, it will increase the speed of your computer or the reliability of data storage.

A RAID array works as follows: a special controller controls a set of hard drives, which are one logical drive. Recording/playback operations are performed in parallel, which ensures high performance. All records are duplicated and checksums are created, this increases the reliability of data storage.

There are several models of RAID arrays.

RAID 0 - high-performance disk array with striping, without fault tolerance;

RAID 1 - mirrored disk array;

RAID 2 is reserved for arrays that use Hamming code;

RAID 3 and 4 - striped disk arrays with a dedicated parity disk;

RAID 5 - striped disk array with “non-dedicated parity disk”;

RAID 6 is a striped disk array that uses two checksums, calculated in two independent ways;

RAID 10 - RAID 0 array consisting of RAID 1 arrays;

RAID 50 - RAID 0 array consisting of RAID 5 arrays;

RAID 60 is a RAID 0 array consisting of RAID 6 arrays.

22.02.2009, 01:30

The essence of the problem: there is a RAID0 array. The disks are only 3 months old and one of them has continuous errors. After tests, I was convinced that it was faulty and intend to replace it with the seller. But because There’s a lot of stuff there, reinstalling which can take up to a month, as well as some information that I didn’t have time to save on DVD, so a corresponding question arose: can I somehow make a backup (or image) from it using data recovery programs? , which can then be transferred to a new disk, so that the RAID array becomes operational again? Thank you in advance for your response!

Added after 31 minutes
In short, so far I’ve found a way to do this using Runtime’s RAID Reconstructor and Captain Nemo Pro. If everything works out, I’ll write, but if not, I’ll continue to ask for advice.

22.02.2009, 11:49

Alexsan, generally speaking, a lot here depends on the method of organizing RAID and the equipment used.

But for fault-tolerant (RAID-1, RAID-5, RAID-10) configurations on normal hardware controllers, it is possible to “hot” replace a disk without additional actions with automatic reconfiguration of the array (even stopping the server operation is not required).

23.02.2009, 03:14

I'm afraid there won't be enough disk space for a full backup. I have two disks with a 600 gigabyte system and two disks with programs and data with a total volume of 1.5 terabytes. So one of the large disks was covered. I'll start trying to restore it tomorrow. While the disks are in the RAID configuration, Runtime's RAID Reconstructor program does not see them. So I’m thinking what’s the best way to start the operation - just cut off one of the disks (normal), or destroy the RAID array through the menu? In the second case, I’m afraid problems may arise with its recovery, because in the RAID menu there is no option to split an array of only two disks and if you use it, the array of normal system disks will also be destroyed. So the question is: what is the best way to prepare for the recovery operation - to mechanically disconnect the disk so that it starts to be read as. IDE or temporarily partition the RAID array via the RAID menu?

23.02.2009, 09:41

Alexsan, I repeat once again -
The correct way is to backup the entire RAID, rebuild the array after replacing the disk, and restore the data from the backup.

there is not enough disk space for a full backup. Either this data is necessary and there will be space (1.5 TB hard drives are on sale), or you are very likely to lose it.

23.02.2009, 12:22

Alexsan, I repeat once again -

In any other case, I personally very much doubt the possibility of successfully completing the operation of replacing a hard drive in RAID-0.
Either this data is necessary and there is a place (1.5 TB hard drives are on sale), or you are very likely to lose it.
As I have written many times (and not just me), RAID-0 is a rare dirty trick, and you can’t store anything valuable on it...

Due to a defect in the disk, it is not possible to make a complete backup for one reason - Runtime's RAID Reconstructor does not see these disks if the RAID function is activated in the BIOS. It turns out that you can only read each disk separately when you set the IDE mode - in this case they are available. I'll try to backup the defective disk on another computer. By the way, I read that the cause of the error may also not be physical damage to the disk, but erroneous writing of data to prohibited service sectors. So after the backup I'll try zero formatting and re-testing. True, I have some questions about this. I have a negative experience from five years ago - a defective Maxtor disk passed all tests without errors after such formatting, but it worked for no more than a week before the first error. Still, I had to send it to Ireland to the manufacturer to replace it with a new one.

© Andrey Egorov, 2005. TIM company.

Modern business in any field is about receiving, storing, processing and issuing information. The volume of such information in an organization doubles on average every year. Unique information is becoming more and more expensive, and its storage requires considerable costs.

RAID array and backup

Let's name industries in which the loss of working information can be literally catastrophic: the armed forces, energy, all types of transport, the financial system, government planning, healthcare, housing and communal services, scientific research...

Two main categories of reasons why information disappears are traditional not only for Russia: roads and fools. Seriously, these are natural causes and the human factor. The first category includes: floods and fires, earthquakes and hurricanes, power outages and hardware problems, i.e. hardware failure for any reason.

The anthropogenic category includes deliberate and “accidental” threats from people: hacker attacks and computer viruses, sabotage and sabotage, software errors and unintentional destruction of data. Trivial human errors account for the lion's share of all cases of data loss: forgotten user passwords and lost encryption keys, incorrect version synchronization, or accidental deletion of a file by the user.

Anything can happen - here's an example. The employee's son, a student, came in to print out his coursework. I sat down at my dad’s computer, began to format the floppy disk (I recently learned), and, of course, by accident, I deleted all the files from the network drive. The existence of a company that has suddenly lost its customer base, history of mutual settlements or other information necessary for its life becomes very problematic...

Preserving important business information is our common goal. In the end, her loss personally should worry the boss much more, because, in the worst case, the system administrator will find another job, and the manager risks not only his information - his enterprise and business, his reputation and fortune, well-being and even health!

The last line of defense against data loss is backup(in English - Backup). That’s why it’s called backup because redundant copies of files and directories are saved to removable media just “just in case.” As we have already seen, data can be lost; and in order to restore them, a backup copy is used. Backup(saving) should be done daily - this copies all new or changed files, so that they are likely to be available for recovery.

In words " backup“There is a whole science involved; practically, this is a real branch of the information business. This concept includes methodology, specialized hardware and software.

For Reserve copy For information storage, primarily tape drives are used, and less commonly, magneto-optical disks, rewritable CDs, or network hard drive arrays. The simplest programs Reserve copy are built into any operating system, but all the wealth of possibilities is provided by commercial software products, for example, BakBone NetVault, Veritas BackupExec, ArcServe, Yosemite TapeWare and others.

All operations Reserve copy automatically appear in the appropriate journal. However, reports on all important operations should be printed, because in the event of data loss, it will be impossible to access the lists of reserved files.

An emergency plan is created in advance. It clearly states who takes what actions in certain cases. All the necessary installation disks are collected so that at a critical moment you don’t have to look for them throughout the office. Detailed instructions are written, which literally outline the procedure step by step: insert CD No. 1, reboot, after the appropriate request, use the magnetic tape (tapes) in accordance with the magazine Reserve copy.

In information theory, indicators of information quality are its relevance and accessibility. Relevance determines the degree of preservation of the value of information for management at the time of its use and depends on the dynamics of changes in its characteristics and on the time interval that has passed since the occurrence of this information. Availability determines the proportion of time during which information is ready for use and is expressed as a percentage: for example, 99.99% (“four nines”) means that no more than 53 minutes of downtime of the information system for any reason are allowed during the year.

To ensure acceptable relevance of information, it is necessary to carry out regular (daily) backup. Availability depends on the speed of recovery of information in case of its loss from backup media.

Tape drives (streamers) were originally created for storing data. They use removable cartridge media, have the highest capacity, have the highest speed, are extremely reliable, are the most cost-effective, provide information security and are an open industry standard. Since its inception, magnetic tapes have gone through five generations of development, have proven their advantage in practice and are rightfully a fundamental element of backup practice.

Recently, it has become fashionable to discuss backup technologies from hard drive to hard drive (disk-to-disk backup, D2D), using RAID array. The prerequisites for this trend, on the one hand, are the emergence of fast and inexpensive Serial ATA hard drives of very large capacity, and on the other hand, the need to carry out the procedure for storing the required amount of data in a limited time. At the same time, SCSI RAID is recommended where the main requirement is the highest speed of storage and recovery, and SATA RAID is used where the ratio of capacity and cost is more important.

The time required to perform a backup is called the copy window. The copy window is determined by the following criteria: Since creating a backup copy requires absolute access to data, this process is carried out during non-working hours, when it has the least possible impact on staff work, server load and local network. Let’s assume that the most ardent workaholics are unlikely to stay up past midnight, and “larks” will not arrive at work before 6 am. Therefore, we get 6 hours, during which all the necessary information must be moved to the backup media - this is where we may need such an important feature of hard drives as their excellent speed.

If the volume of data transferred during one session exceeds the size of the magnetic tape cartridge, it becomes necessary to manually replace the media at the request of the program Reserve copy. Obviously, this can be problematic at night (by the way, this is why the most prudent system administrators choose not a simple drive, but an autoloader or tape library). In addition to excellent speed, a hard drive storage system can help us in this case by providing free space for recording a backup copy that exceeds the capacity of a single cartridge.

The advantages of storing backup copies of data on hard drives (primarily high speed of storage and recovery) and on magnetic tapes (low cost of storage and unlimited capacity) are combined in disk-to-disk-to-tape (D2D2T) solutions. This approach involves using the disk cache as an intermediate step in the procedure Reserve copy, whose ultimate goal is still magnetic tapes.

Andrey Egorov, head of the corporate client department at TIM, certified professional - MCSE, Master CNE, CIA, ICIS, etc.