Types of RMAN Backups. Full, incremental and differential backups Differential backups

Incremental backup allows you to effectively save information that is constantly changing: documents, projects in development, mail backup, etc. Handy Backup - incremental backup program any files.

What is an incremental backup?

Incremental copy Is a copying method in which additions are added step by step to the original copy of a dataset to reflect changes in the data (these stepwise changes in the dataset are called increments).

For example, if out of 200 files in the original set, only 3 are changed, then they will be copied during the next incremental backup.

An incremental backup program should provide the ability to regularly make such additions, as well as to restore them step by step if necessary.

In this mode, the incremental backup is repeated cyclically, starting with a complete data backup. When the number of incremental backups reaches the specified value or age, the whole cycle repeats automatically, starting with the creation of a new full copy.

The key advantage of a mixed incremental backup is its ability to break off long chains of increments. This can slightly slow down the backup process, but makes data recovery more reliable and faster.

How to make an incremental backup of files in Handy Backup?

It is very easy to program an incremental backup task in Handy Backup. Select in Step 4 in advanced mode * of task creation incremental or mixed incremental copying.

For mixed incremental copy also set the length or repetition time of the chain of incremental backups in the version settings dialog.

* At Step 1 of creating a task, you must check the box next to the "Advanced mode" item.

Recommended solution for incremental backup

Cross-platform incremental backup solution

Incremental copy of files and folders in Linux and over the network

In addition to the Windows version, Handy Backup also fully supports Linux distributions based on Ubuntu 16.04 and 14.04 at the executable level. It also provides a Java workstation for networked Windows, Linux and FreeBSD machines.

Try Handy Backup's capabilities for incremental file backup yourself,
by downloading and installing a free 30-day trial version of the program with all the functions!

Backing up data is something that every active computer user who does not want to lose all of his information (or part of it) in case of an unexpected failure, should regularly perform. Often in various applications designed for backing up information, you can find three mechanisms for how to create a copy: full, incremental or differential. In this article, we will consider how these copying methods differ from each other.

Table of contents:

Data backup methods

There are many programs that are designed to create a backup copy of information, both in the Windows operating system, there and in the Mac OS. They all perform approximately the same actions - create a backup copy of the operating system, completely copy the disk, some of its partitions, folders or other data, depending on the settings selected by the user. Then these backups can be used to restore information.

The created backup copy needs constant updating. Based on the conditions for creating a backup used in the program, you can create a copy, while choosing a backup mechanism:

• Full copy creation;
• Generating an incremental copy;
• Create a differential copy.

These actions are available in many applications, for example, in one of the most popular data backup software, AOMEI Backupper. Within the framework of this article, examples will be considered on it, but you can find similar backup mechanisms in other programs.

Full backup

With this method of backing up, system snapshots that are generated as part of a single backup task can work independently of each other. Damaging one of these images will not affect the performance of the others. That is, with a full backup, the system snapshot contains all the information being backed up.

The full backup method is the most reliable, but also the most wasteful in terms of resources. It will take tens of gigabytes to back up the Windows operating system and a few small applications. Accordingly, it is irrational and wasteful in terms of free space on the drive to constantly save such full-fledged backups and store them on the hard disk. This is why two other mechanisms are used, discussed below.

Incremental backup

Incremental data backup implies that the user, when creating a backup, once generates a full copy of the system and all files, and all copies created in the future are children of the main and previous ones, that is, they contain only information about the changes that have occurred - deleted, modified and created files.

Thus, each subsequent incremental copy after the first contains only information about the changes. It looks like this:

• Second copy. Child - contains information about changes in data since the creation of the first copy;
• Third copy. Child to the second - contains information about data changes since the second copy was created.

A plus of this method of backing up data, in comparison with the first, is a smaller copy size (each new incremental copy weighs tens to hundreds of megabytes, depending on the number of changes that have occurred). Minus - the reference of each new copy to the previous one during recovery. That is, if one of the copies is damaged, you will have to restore to the last working copy in a continuous chain from the first. In addition, recovery from an incremental backup takes longer than other backup methods.

Differential backup

The differential copying method is close to the incremental copying method, but there is a key difference between them. As part of differential copying, new images are children of the first one.

This means that during the first backup using the differential method, a complete copy of the system is created, after which all subsequent snapshots contain information about the changes that have occurred from the first copy. It looks something like this:

• First copy. The main one contains all the information;
• Second copy. Subsidiary - contains information about changes in data since the creation of the first copy;
• Third copy. Child - contains information about changes in data since the creation of the first copy.

As you can see, the third copy with the differential backup method is not a child of the second. That is, if problems arise with one of the differential snapshots, you can recover to any other working differential copy. This is the key difference between differential and incremental backups.

Each differential snapshot is larger than the incremental snapshot because it needs to store information about all changes since the first full copy was taken. Moreover, each new differential image will weigh more than the previous one.

Which backup method is best

Having considered three backup methods, each user can independently conclude which option is best for him. Let's briefly summarize and present several scenarios:

• Full backup. The most reliable way. Suitable for those users who have the ability to store large-volume backups;
• Incremental backups. Best option for users who back up to a small disk, such as an SSD. The advantage of this method, in comparison with differential backup, is only in the size of each new system snapshot;
• Differential backup. Best option for home computer users. With this method of copying, you only need to attend to the safety of the first copy.

There has been a lot of talk and writing about backup lately. And we, SIM-Networks, as well. :)

The fashionable theme is inevitably mythologized. We tend to fill in the gaps in our knowledge with fictitious facts and subjective assessments. This is the case, in particular, with regard to the backup service and the issue of its organization by hosting providers. Should the hoster provide its customers with backups automatically, by default? The answer to this question can be found in our material.

The increased interest in the topic of backups is not surprising: given the active development of malware, the outstripping development of antiviruses, it is most rational to build IT security around a backup system - instead of spending resources on preventing attacks and fighting viruses, it is much easier, cheaper and easier to raise the system and saved data from current backups.

In addition, an up-to-date backup will help neutralize the consequences of force majeure or human factors, as well as equipment failure due to various reasons. It is not for nothing that one of the commandments of the system administrator says: when preparing a new server for work, first set up a backup!

How to set up a backup

You can make a backup yourself - there are enough tools today, Google will gladly tell you. But if you are not a tough pro in the field of system administration, it is better to trust those who are competent and able to set up a backup, fully responsible for the result.

It is very important to pay attention to two points: copies of information that are critical for you should be made regularly, and stored in a remote place, as far as possible from the originals.

The first point is important because the information at the time of recovery should be as relevant to you as possible. For example, if your system is infected with a virus and the only way to recover valuable data is to restore it from a backup, then, you must agree, it will be very disappointing if the most recent copy of your financial statements is dated last month.

The importance of the second point can be illustrated as follows: if your backup storage for backup copies is located on the same server as the main system, then if the server burns out, everything will really burn. Finally and irrevocably.

Therefore, we take care of the correct schedule of backups and ensure the remoteness of storage for copies.

The main criteria for choosing a program for backups

In the event that you still want to take the risk and start organizing backup of your data yourself, experts recommend that you be guided by four universal criteria when looking for a backup program:

• resource efficiency: the program should work in the most autonomous mode (without distracting you and without wasting your time resource, that is, automated as much as possible), with the minimum possible load on system resources and executed in the minimum possible time;
• recovery rate: The software should restore your data from a backup copy as quickly as possible so that business processes do not suffer; the function of working directly with copies of data would be ideal;
• data protection and security: the backup program must provide you with a sufficient level of security - both cryptographic and hardware (protection of data transmission channels in the storage system, data protection during a backup operation, the ability to restore an interrupted session)
• flexibility: The software should be equally suitable for all types of data (since it is impossible to predict which of them you consider critical and choose to be copied to the backup storage system), as well as give you the opportunity to choose backup methods and function equally fully with any of them.

Modern software used by professional admins always meets these criteria. In addition, people who are specially trained and have a rich and varied experience in setting up a backup can choose the most optimal backup option for each specific case. Therefore, we still strongly recommend that you seek help from specialists, so that later it will not be excruciatingly painful from overwritten correct copies, over which erroneous information is written. It is clear that restoring such versions of backups will not bring you the desired result, because the original correct data has been lost. This happens if the wrong copy method is selected and the backup storage capacity is too small.

Now let's talk about the types of backups - full, incremental and differential. They differ in the way they copy and compress information.

Full backup

Everything is clear from the name: every time, according to the backup task, a complete copy of the entire system is created, more precisely, all the data that you specified for backup when setting the task. To reduce the total volume of the backup, all data is compressed into an archive. Thus, during a full backup, archives appear in your storage at a predetermined frequency, where the data are mostly duplicated (since they do not change for a long time). This is a serious drawback, because a huge amount of resources is consumed (see item 1 in the list of backup criteria): storage space, creation time and processor time, computing power, and finally, traffic resources when transporting archives to a remote storage system. And although the method of full copying was previously very common due to its high reliability, today it is recognized as ineffective in its pure form. For example, for backups of low depth (less than two weeks) or with high frequency (once a day, once every few hours), a full backup is overwhelming.

The mechanism will save the situation a little deduplication - identification and removal of duplicate data in full copies. It is also set by special software both at the storage or server level, and directly on the client. Statistics in some sources give impressive results for deduplication rates - from 90% to 98%.

The advantage of a full backup can be called only the speed of recovery: when data is raised from one archive, it happens faster than with an incremental or differential backup.

Today, the full backup method is usually used exclusively as a basic one in combination with other methods that are less resource intensive. Sometimes this approach is also called mixed or synthetic backup.

Incremental backup

Compared to full backup, it is much more economical and faster, as this process only backs up files that have changed since the previous backup. The original data recorded initially is not overwritten. The mechanism of incremental copying is simple: the time (for example, midnight from Sunday to Monday) is selected as the starting point of the backup X 0, at which the full backup is made; at point X 1 (midnight from Monday to Tuesday), files are copied that have been modified and / or appeared since the moment X 0; at point X 2 (midnight from Tuesday to Wednesday), files that have changed / appeared since the execution of X 1 are copied; ... at point X n the cycle ends and the next full backup is made.

This method is much more economical in using resources and storage space, and time and data transfer traffic, compared to others. However, when restoring data, if necessary, from a backup copy, a step-by-step recovery from points X n-1 ... X 2, X 1, X 0 - up to and including the last full backup, and this process can take a long time.

Differential backup

It wins over the incremental one in the case of data recovery - it has less time for this operation, since full copies of X 0 and X n are compared and no step-by-step recovery is required. However, in terms of the amount of space for placement in the storage system, differential backup is comparable to full, therefore, saving space in storage and traffic is practically not achieved.

With a differential backup, a "cumulative total" copying occurs: each changed file is copied again at each subsequent backup point. That is, it looks like: X 0, X 1, X 1 + X 2, X 1 + X 2 + X 3, ... + X n, X 0 + X (1 + ... n)

In a word, it is very cumbersome and difficult when calculating storage space.

Difference between incremental and differential backups is easy to understand. In fact - it is in one word. Just compare:

• incremental processes files modified or created since the last backup;
• differential processes files modified or created since the previous execution complete backup.

Other types of backup

A kind of differential backup is delta copy (delta blockor delta style backup). With this method, only the changes that occur in the files are written to the copy, and not completely changed data is overwritten. That is, a piece is copied, not the entire file. True, the delta-block method can be applied specifically to the files being modified, and not to the created files - therefore, the new files are copied entirely.

It is distinguished by its high creation speed, extreme space saving and significantly less (compared to incremental and differential backups) amount of redundant data. It would seem that everyone should use the delta, but this does not happen, since the creation of backups in this way and the restoration of information occurs using special software. In addition, recovery from a delta backup takes a very long time: data has to be collected from a mosaic of changed pieces. Nevertheless, this method is convenient to use to ensure continuous data protection (when a file is backed up immediately after it is created or modified - a mechanism that remotely resembles autosaving in Word files))) or in cases of reduced bandwidth when saving backups in a remote storage system.

Similar to a delta-block backup, the developed by programmers works binary patch method, in which parts of the changed files are copied, but a different comparison base is used (in delta - blocks, in this method - bits of information).

However, keep in mind that both of these methods are used in conjunction with differential or incremental backups, but not by themselves.

Backup is sometimes called technology mirroring, used, for example, at the hardware level in RAID1 or when creating mirror sites. In essence, this is a simple copying of the original and modified files, without archiving and systematizing the accumulation of modified files in a given period.

Over the past 12-15 years, many critical changes have taken place in backup technology, forcing us to rethink the effectiveness of approaches and discover new ways. For example, the introduction of technology snapshots (snapshots) - snapshots of the file system, from which you can "glue" the backup, - allow you to make backups in cloud systems quickly and painlessly without stopping the virtual machine. In addition, when used in the cloud, Snapshots can seriously save storage resources, since they do not take up space on the client's disk.

SIM-Networks customers choose backup!

Of course, if you like to do everything yourself, it will not be a problem for you to set up a backup manually - on your home computer. True, even in this case there is a partial risk, because something may go wrong, and valuable photographs, books, videos or calculations of the rocket stage may accidentally not be saved or remain with a defect that will make it impossible to restore them from a backup copy. And if we are talking about office machines? What if you need to provide a backup of the data stored in the corporate infrastructure? We recommend that you rely not on your own strength, but on the professionalism of the hosting provider. Ordering a backup setup and space for remote storage of backups in Germany is very easy.

Backing up your business-critical data has become a necessity these days. And also - part of the comprehensive measures to ensure the information security of the company, which we wrote about in the material

If you are renting capacity in our cloud infrastructure, ordering a SIM-Cloud BaaS backup service is as easy as shelling pears, in a couple of clicks. Everything is already configured and will be connected automatically as soon as you give the command. By the way, when our engineers were developing SIM-Cloud BaaS, they analyzed the effectiveness of different types of backups and opted for the incremental backup method. Our cloud backups are optimized to average 15 to 30 minutes RTO (time to restore data from a copy), depending on the amount of data. Cloud BaaS from SIM-Networks meets all of the above criteria for high-quality backup.

You can independently choose in which data center to organize a storage for backups. The first option is local storage: your backups are stored in the same DC where your main infrastructure is deployed. This makes it possible to speed up RTO and RPO. The second option is that backups are sent for storage to a data center remote from the one in which the main infrastructure is deployed. Data recovery in this case will be a little slower, but the safety factor is higher. If you are in doubt about which option to choose, please contact our Customer Care team to help you find the best solution.

And for adherents of classic "hardware" we offer rent for backups: reliable, safe, high-tech. And, of course, our highly qualified support experts will help you set the required frequency, depth and other parameters for backing up your system.

Greetings, dear site visitors! Let's continue the topic started in the last post of this heading and this time we will take a closer look at how incremental backups are carried out.

Each data block in a data file contains a system change number (SCN), which is the SCN at which a new change was made to the block. During an incremental backup, RMAN reads the SCN of each data block in the input file and compares it to the SCN checkpoint of the parent incremental backup. If the SCN in the input data block is greater than or equal to the SCN of the parent's checkpoint, then RMAN copies the block.

Note that if you enable the block change tracking option, RMAN can refer to the change tracking file to identify changed blocks in the data files without scanning the entire contents of the data file. Once enabled, block change tracking will not affect how you perform or use incremental backups, but only offers improved performance.

Level 0 and Level 1 Incremental Backups

Incremental backups can be level 0 or level 1. An incremental backup of Level 0, which is the basis for subsequent incremental backups, copies all blocks containing data, backing up the data file to a backup set, just like a full backup. The only difference between a level 0 incremental backup and a full backup is that a full backup is never included in the incremental strategy.

Level 1 incremental backups can be of one of the following types:

• Differential backup, which backs up all blocks changed since the last incremental backup at level 1 or 0
• Cumulative backup, which backs up all blocks changed since the last incremental backup at level 0

Incremental backups are differential by default.

The size of the backup file depends solely on the number of modified blocks and the level of the incremental backup.

Differential Incremental Backups

In a level 1 differential backup, RMAN backs up all blocks that have changed since the last cumulative or differential incremental backup at level 1 or 0. RMAN determines which level 1 backup was the last time and backs up all blocks modified after that backup. If no level 1 backup is available, RMAN copies all blocks changed since the level 0 backup.

The following command performs a level 1 differential incremental backup of the database:

RMAN\u003e BACKUP INCREMENTAL LEVEL 1 DATABASE;

If a level 0 backup is not available, then the behavior depends on the compatibility mode setting. If compatibility is\u003e \u003d 10.0.0, RMAN copies all blocks that have changed since the file was created and saves the results as a level 1 backup. In other words, during an incremental backup, the SCN is taken equal to the SCN of the file creation. If compatibility<10.0.0, RMAN генерирует бэкап уровня 0 содержимого файла во время резервного копирования, чтобы не было противоречия с предыдущими релизами.

Figure 1 Differential Incremental Backups (default)

• On Sunday
  Level 0 incremental backup backups all
• Monday through Saturday
  Every day from Monday to Saturday, a differential incremental backup of level 1 backs up all blocks that have changed since the last incremental backup at level 1 or 0. For example, a backup on Monday copies blocks that have been changed since Sunday's backup of level 0, a backup on Tuesday copies blocks changed by starting a level 1 backup on Monday 1, etc.

Cumulative Incremental Backups

In a cumulative level 1 backup, RMAN backs up all the blocks that have been used since the last level 0 incremental backup. Cumulative incremental backups reduce the work required for restores, ensuring that you only need one incremental backup from any given level. However, cumulative backups require more space and time than differential backups because they copy work done by previous backups at the same level.

The following command performs a cumulative level 1 backup of the database:

BACKUP INCREMENTAL LEVEL 1 CUMULATIVE DATABASE; # blocks changed from level 0

Figure 2 Cumulative Incremental Backups

In the example shown in, the following happens:

• On Sunday
  Level 0 incremental backup backups all blocks that have ever been used in this database.
• Monday through Saturday
  A level 1 cumulative incremental backup copies all blocks changed since the last level 0 backup. Since the last level 0 backup was done on Sunday, a level 1 backup every day from Monday to Saturday backs up all the blocks that have changed since Sunday.
• The cycle repeats over the next week.

Simple Incremental Backup Strategy

Select a backup scheme according to an acceptable MTTR (mean time to recover). For example, you can implement a three-tier backup scheme so that a full or level 0 backup is taken monthly, a cumulative level 1 backup is taken weekly, and a level 1 differential backup is taken daily. With this scheme, you will never need to use more than one day's supply of transaction logs for full recovery.

When deciding how often to take a full or level 0 backup, use a good rule of thumb: you should take a new level 0 backup every time 50% or more of the data has changed. If the rate of change of your database is predictable, you can monitor the size of the incremental backups to determine when to take the next level 0 backup. The next query displays the number of blocks written to the backup set for each data file with at least 50% its reserved blocks:

SELECT FILE #, INCREMENTAL_LEVEL, COMPLETION_TIME, BLOCKS, DATAFILE_BLOCKS FROM V $ BACKUP_DATAFILE WHERE INCREMENTAL_LEVEL\u003e 0 AND BLOCKS / DATAFILE_BLOCKS\u003e .5 ORDER BY COMPLETION_TIME;

Compare the number of blocks in differential or cumulative backups with the base level 0 backup. For example, if you only create cumulative level 1 backups, then after taking another new level 1 backup with a size of approximately half the size of the base level 0 backup, take a new level backup 0.

Thanks for your attention!.

An incremental ("incremental") backup copies only files that have changed since the last time a full or incremental backup was performed. Subsequent incremental backups only add files that have changed since the previous one. On average, an incremental backup takes less time because fewer files are backed up. However, the data recovery process takes longer as the data from the last full backup plus the data of all subsequent incremental backups must be restored. At the same time, unlike differential copying, changed or new files do not replace old ones, but are added to the media independently.

Cloning

Cloning allows you to copy an entire partition or medium (device) with all files and directories to another partition or to another medium. If the partition is bootable, then the cloned partition will also be bootable.

Image backup

An image is an exact copy of an entire partition or medium (device) stored in one file.

Real-time backup

Real-time backup allows you to create copies of files, directories and volumes without interruption, without restarting your computer.

Rotation schemes.

Changing the working set of media during the copy process is called rotation. For backups, a very important issue is the choice of a suitable rotation scheme for media (for example, tape).

One-time copying is the simplest scheme that does not provide for media rotation. All operations are carried out manually. Before copying, the administrator sets the start time for the backup, lists the file systems or directories that need to be backed up. This information can be saved in the database so that it can be used again. For one-time copying, full copying is most often used.

Simple Rotation Simple rotation implies that a set of tapes are used cyclically. For example, a rotation cycle might be a week, in which case a separate media is allocated for a specific workday of the week. The disadvantage of this scheme is that it is not very suitable for archiving, since the number of media in the archive is growing rapidly. In addition, incremental / differential recordings are performed on the same media, which leads to significant wear and tear and, as a result, increases the likelihood of failure.

"Grandfather, father, son" This scheme has a hierarchical structure and involves the use of a set of three sets of media. Once a week, a full copy of the computer disks is made ( "father"), incremental (or differential) copying ( "son"). Additionally, once a month, another full copy is carried out ( "grandfather"). The composition of the daily and weekly set is constant. Thus, compared to simple rotation, the archive contains only monthly copies plus the latest weekly and daily copies. The disadvantage of this scheme is that only the data that was available at the end of the month, as well as the wear of the media, are included in the archive.

Tower of Hanoi The scheme is intended to address some of the shortcomings of the simple rotation and rotation scheme “Grandfather, Father, Son”. The scheme is based on the use of several sets of media. Each set is designed for weekly copying, as in a simple rotation scheme, but without the removal of full copies. In other words, a separate set includes media with a full weekly copy and media with daily incremental (differential) copies. A specific problem with the Tower of Hanoi scheme is its higher complexity than other schemes.

"10 sets" This diagram is for ten sets of media. The forty-week period is divided into ten cycles. During the cycle, each set is assigned one day of the week. After the four week cycle, the set number is shifted by one day. In other words, if in the first cycle, set number 1 was responsible for Monday, and number 2 for Tuesday, then in the second cycle, set number 2 was responsible for Monday, and number 3 for Tuesday. Such a scheme allows evenly distributing the load wear and tear between all media.

The Tower of Hanoi and 10 Sets schemas are not often used because many backup systems do not support them.

Backup storage

1. Streamer tape - recording backup data onto the tape of the streamer;

2. "Cloud" backup "- recording backup data using" cloud "technology through the online services of special providers;

3. DVD or CD - write backup data to compact discs;

4. HDD - recording backup data to the hard disk of the computer;

5. LAN - recording backup data to any machine within the local network;

6. FTP - recording backup data to FTP servers;

7. USB - write backup data to any USB-compatible device (such as a flash card or external hard drive);

8. ZIP, JAZ, MO - backup to diskettes ZIP, JAZ, MO.