In this document I will discuss about the theory of backups. I will analyze such questions as: What is a backup? Do I need to backup all my systems files or only my data? Do I need compression? Do I need to use Volume Shadow Copies? …and other important questions.
Everything expressed here is my opinion only and should not be taken as “the one and only truth”. This is the way I make my backups, and how my mind works when I’m planning them.
What is a backup and why do I need to backup my files?
Did you know...?
There are more than 40000 threads there at this moment.
In the computer world, a backup is a copy of some data. This copy could be used to restore the original data when the original information is lost or damaged. You can make backups of your data manually, by copying your files to another place: a CD, another disc, another machine, to a tape device, etc, or it can be made automatically, by scheduling a backup. Ideally, the copy should be stored on another physical place and should not be stored on the same room where the original is. In case of disaster, like a fire, having both the original data and the backup on the same physical place could be fatal. I recommend you to make multiple copies of your valuable data: for example, you can have a copy stored on another hard drive and another copy on some remote FTP server, for maximum security.
Do I need to make a backup? Why do I need Cobian Backup?
Yes, you need to make backups of your files, unless you don’t care about your data at all. Do you need Cobian Backup? Actually not at all. You can make your backups manually, but it’s very easy to forget to backup your files regularly. Your data is often not static, it may change often and you need to be careful and backup any changes you do with your data. %PROGRAMNAME% just helps you to “setup your backup and forget about it”. It will then back up your files for you without any human interaction.
Permissions are the Alpha and the Omega of a modern OS
In a modern multiuser operative system, everything is controlled by access rules (or permissions, for simplicity). Every process that is executed on the computer has an owner, and it can only have access to those resources that the owner is allowed to use.
“But, I am the only user on my computer. I know that” – say many people. Wrong, think again. Every default Windows installation includes many other users that are not you (even if you don’t even think that they exist). For example: the local System account is the default account used to execute Windows Services. Guest, Administrator, ASPNET, HelpAssistant, etc are other accounts that may exist on your computer.
If you are executing the backup program as the user A and want to back up the disc C:\, which includes files which are owned by the user B, it’s very probable that this operation will fail, because A by default shouldn’t have access to B’s files.
“But I’m running the program as an Administrator. I have access to all files”- This is another myth. And sure it seems logical that the almighty Administrator should have access to all files, but… that is not necessarily true. Being an Administrator means that you can potentially access any file but if some user (or the system) eliminates the Administrator as a member of the list of allowed users, guess what… the backup program will not have access either to those files.
In Vista, it's also especially important to know that even when running as an administrator, your program could be executed with limited powers if you have UAC activated. Then we have some other kind of permissions: policies, which can also limit an account in several ways. For example, the local System account cannot access network resources by default.
In my experience, in the 9 years offered support for Cobian Backup, I can tell that more than 70% of the problems that the users experiment are just problems with permissions and could be easily solved if the user only could educate him- or herself a little about how permissions work on a Windows system.
Do I need to backup system files or only my data?
Well, it depends. Some people prefer to back up the whole system (OS plus data files). Yes, today storage is really cheap and all, but … backing up the thousand of files in different states that a modern OS has is not a trivial task, it takes a lot of space, a lot of time, and storing it on a remote machine can be a nightmare for your internet connection. For me: I backup what is REALLY important and irreplaceable: my data files.
The system can be re-installed from the original media. Your programs can be re-installed as well, but… your data? Your documents, spreadsheets, databases, photos…. That cannot be re-created as easy. And THAT is what you need to care about.
Do I need compression? And Encryption?
Compression is a good thing in theory. You have X bytes of data and, voila, now you have the same data, but using less space. There are different methods of compression and all of them have different advantages/disadvantages in relationship to the other. But compression as cool at it seems has actually some disadvantages: all your data is stored in one file, so if only one byte on that file is corrupted, you can lose all your data at once!
Compression is often a slow process, much slower than a plain copy. Updating a large archive takes also a lot of time and sometimes you need a lot of extra-space as well. Compressing directly to a remote location can eat a lot of bandwidth as well. Many files like mp3, jpg, etc compress very badly because they are already compressed, you may be wasting your time compressing them. Oh, and did I say that storage is very cheap these days?
That may sound like I’m against compression. Well, I’m not, really. But I don’t use it for my backups.
Encryption? That depends on your data. You may not need encryption at all. You maybe already have an encrypted volume (NTFS drives can use encryption), or maybe you are using a TrueCrypt volume as your destination. In those cases you don’t need an extra layer of encryption, unless you are *very* paranoid about your data. Sometimes it may be enough to use NTFS file permissions to prevent the access to your backup files. Bear in mind that encryption is also a slow operation.
Do I need to use Volume Shadow Copies?
Not if it’s not necessary. Volume Shadow Copy Services is a technology that can be used to copy files that are locked or in exclusive use by some other application. The creation of the snapshot is pretty fast, but there is no need to force the system to make a shadow copy if you know that none of the files to backup will be in use.
Is it better to create separated backups or to overwrite?
It depends. Sure it is “prettier” to overwrite or update a backup, and have all your files in one place, but having separated backups (full or not) has several advantages: you can maintain several versions of your files, for example. But the more important thing is, that by overwriting an old backup you are taking the risk that if the backup is corrupted, interrupted, or canceled, you may end up having a backup in bad state, so use separated backups when possible and then, use the automatic function for deleting old backups.