Hi Folks, I have three questions.
First, I need a "Theory of Operation" tutorial. Amanda has a lot of moving parts, and I don't have a very clear understanding of what they are or how they work. If I had a better understanding, I might be able to answer a lot of my own questions. For example, If Amanda is backing up to a NAS, can Amanda simply direct that client to that NAS, or must Amanda get in the middle, thereby doubling the network traffic -- one file transfer from the client to Amanda, and a second from Amanda to the NAS? Can Amanda create one vTape per backup, which vTape will be the correct size for that backup or must they all be per-arranged? You get the idea... Second, I need to read tutorials on configuration for disaster recovery, meaning backups that can be "bare-metal" restored -- on both Microsoft Windows and Linux systems? Of course, taking the backup is only half the problem; restoring the backup is equally important... Third, and finally, I've asked this question before and I've been roundly scolded for my ignorance of both theory and practice of Amanda, so I went off to lick my wounds, and defer my goal of implementing an Amanda backup system. I've mostly recovered, so I have returned more determined to see if a simple answer to my question exists. I'm fully aware that most of the collective experience with Amanda is about understanding what Amanda wants to do and what Amanda is doing and how to accommodate Amanda, and not about how to force Amanda to accommodate my limited understanding, but that is the position where I find myself right now. I know Amanda is very good at managing backups to optimize both the time and space necessary to execute a large backup set. Amanda is only one of many things I must understand and monitor. I can't devote as much time to it as I would like. I must force Amanda to do what I expect so I don't have to research and understand Amanda's behavior, and reach a zen level or understanding, which will sublimate over time, only to have to be re-acquired at some sub-optimal time -- during a disaster. I run a backup every night. My other backups are differential filesystem Sunday night through Thursday night, full filesystem on Friday night, and disk image on Saturday night. I need Amanda to conform to this schedule. If I ever grow to where I am backing up many systems and time and space are critical resources, then I will have justification to spend more time and have more faith in Amanda, but for now my own bandwidth is the critical resource. Please don't scold me because I don't understand that Amanda is very good at doing what I am preventing Amanda from doing and I should simply let Amanda do what Amanda does best; I know that, but allowing Amanda to do this means that I have to understand two different backup paradigms. Please don't scold me because I could simply use tar or some other technique and write my own scripts. I know this as well, but that small development requires debugging and Amanda already has this stuff debugged. I don't want to discover that my scripts have a bug at some sub-optimal time -- during a disaster. Thanks for the help, -- Chris. V:916.974.0424 F:916.974.0428
