What is backup?
Documents are very important in corporations and businesses, as we already know. There are many types of computer generated documents like these: invoices, product offers, brochures, resumes, project proposals, business letters.
For administrating and saving these documents in order to use them for different projects, different types of document management software, billing systems, financial software, CRM software are invented. All of these software generate a lot of corporate documents on daily basis.
The problem starts when some of these documents are lost because of hardware faults.
How can this document and data loss and corruption be prevented?
We need to make copies of all corporate data and documents on different physical locations and this process is known as Backup.
There are different types of backup processes and there are many software for making this process easier.
One way to categorize backup processes is:
- Backup a whole server – Being able to recover and restore a working environment, with the data included
- Backup of files – Having your data in a separate location so you can restore if it gets lost or compromised
Backing up a whole server
In Linux everything is file. So copying and restoring all of the files on the root “/” filesystem and below, including the mount-points, would clone server completely.
However, if you copy to a different hardware configuration, different drivers and disk configuration might be needed for the system to work properly. This is a completely separate issue and will be covered in one of our future blogs – “Hot-clonning a running Linux system”.
Today however, virtualization and “cloud” severs are used more often and it’s rather simpler to copy a whole virtual server, instead of a physical one. This blog will focus more on file-based backup, so strategies for cloning a whole server will be covered in one of our future blogs.
Backing up only the files is not something which can be done efficiently when backing up a whole server. Because with files you have different versions of the files and if you want to restore a file from one week, you don’t need to restore the whole server. Also, backing up files, takes far less disk space instead of creating a snapshot or an image of your running server.
For file-based backups we at VapourApps use BackupPC, as it is a disk-to-disk, simple and efficient and we’ve made it a part of our App Suite, integrating it with our Owncloud, Filesharing and Email applications.
You can find how to install and configure Backuppc here.
We recommend configuring:
- Full backup once a week
- Incremental backup every day or every 6 hours, depending on how often the files change within a company.
On the image above, you can see a summary of the backup of five different hosts, the date of the last backup, its size and how many full backups are stored historically.
One can also display a history of backups for a particular host, in this case 192.168.100.100. The backups are listed, the duration which is needed for each backup and whether it is a full backup or an incremental one. If you click on the particular backup, you can choose to restore all or only some of the files, to the version which was copied at that particular backup. This can be useful if a file has been corrupted and you don’t remember when was the last good version of it. You can return multiple versions from different dates, for a single file.
A graphical display of the storage needed to support the current backup strategy is also displayed, where one can see that the incremental backup only increases the used storage by a small amount, since it just copies the modified files to the backup host.
Backuppc is very useful, since it can backup your files from servers, but also from your workstations. And it doesn’t require additional licensing. What you need is a large enough disk to store the data.