Answer the question
In order to leave comments, you need to log in
What backup system should I use for *nix?
Good afternoon, ladies and gentlemen!
Given: linux server farm, with web servers, databases (mysql, postgres), file dumps, etc. everything is quite standard.
Task: back up all this to the disks of the backup server and from there transfer it to one more place.
In fact, the task is not difficult and has already been implemented, but with self-written scripts. I would like to find a replacement for them in the form of an adequate product.
The requirements for the product are as follows: a client-server model, the ability to make full / incremental copies, the presence of a web interface, at least in order to visually see whether all backups are going through correctly.
From the googled options, I looked at bacula, backuppc, amanda. Most of all I liked the first option, well, the bacula has an incomprehensible backup rotation system, and it’s not entirely clear what to do with ready-made backups if the server itself with the bacula dies. Backuppc disappeared due to the lack of a client and the need to fumble everything via smb, but amanda seemed redundant, although I confess I looked at her the least.
Answer the question
In order to leave comments, you need to log in
Worked with Bacula and Amanda. Since no tape devices are used, they are redundant in your case. I would write a script using rsync, for files and web. Also backup bases means of a DBMS. To track the web server is not the most convenient solution, I came to the conclusion that after a while you stop visiting it. I tried to send myself a full log by mail, the effect is the same. Now the script sends me one message a day, to mail/jabber something like this:
gdi COMPLETED ARCHIVELOG 04-02-2014 07:00 - 04-02-2014 07:01
nod COMPLETED ARCHIVELOG 04-02-2014 08:00 - 04 -02-2014 08:00
The word Failed catches my eye, then I look at the logs.
I have been using bakula for a long time and not in one place.
The system is just class.
It meets all the requirements, it even has a web interface, although I never installed it because I didn’t need it. Once I set it up and forgot about backups, the status of tasks and logs come to the mail, well, certain parameters are also monitored by Zabbix.
With the loss of the bacula database and the safety of archives, it is not difficult to restore data, but for a long time, this is probably the biggest disadvantage of this system. At some sites, bacula is also used for backups from servers running Windows. There are no complaints either, the only thing I encountered was that when restoring the directory with the restored data is not visible through the explorer, since the system and hidden attributes are set to it during restoration. Well, simultaneous removal of two copies through VSS is impossible due to the limitations of bacula, which, when using VSS, sets a semaphore in which the bacula client cannot create more than one VSS volume for all disks.
Also, with the help of Bakula, I implemented a backup system according to the disk-to-disk-to-tape scheme.
If there are more than 5 servers and a lot of data that needs to be backed up according to different schemes and stored for different amounts of time, then Bakula definitely rules.
csync for the file system, for pg you can also do streaming replication or configure archive_command, for mysql extra backup.
I can recommend using backupninja for incremental backups.
The main advantage is a convenient and simple implementation. There is no web interface, but all configuration is specified through a pseudo-interface in the console, so you don’t have to write it manually.
It uses rsync for incremental backups, it is possible to export databases, store backups both locally and remotely. Possibility for enciphering is implemented.
If you need to do something yourself, then you can easily write a plugin for it.
Didn't find what you were looking for?
Ask your questionAsk a Question
731 491 924 answers to any question