Building right system for my needs [closed] – Managing your servers can streamline the performance of your team by allowing them to complete complex tasks faster. Plus, it can enable them to detect problems early on before they get out of hand and compromise your business. As a result, the risk of experiencing operational setbacks is drastically lower.
But the only way to make the most of your server management is to perform it correctly. And to help you do so, this article will share nine tips on improving your server management and fix some problem about linux, failover, cluster, disk-space-utilization, .
I will be providing “File-Hosting” service and I need something that works with me along the line and allow the project to expand.
At first I was thinking about one big server with RAID 5 as a start-up, but then I thought
using multiple servers would be better as it’s easier to expand in the future and should be cheaper, but I have absolutely no idea how these servers should/can be linked and if multiple servers would act as a single system.
Can you help me out with ideas and links for information on linking multiple servers together to make them act as one system so I can use all processors & disk space as if it was one server + if one server crashed what are the options available to recover its data? and will the other servers keep working normally?
I still have more questions about geographically different location for servers, and how to pass files from one location to another but it’s not a priority for now.
I’ve read about this feature on Google but they’re using their own file-system and other things that are not available for public.
Start off by looking at publicly available distributed filesystems such as GlusterFS. Systems like that can provide you with redundancy and scaling.
As for sharing CPU power, you can load balance requests among multiple machines depending on how you’re serving files. Tools like HAProxy can help with this.
In this setup, you’re linking your filesystem with redundant copies of files so that if a node goes offline you don’t lose availability of files. Similarly, HAProxy can make several different servers appear as one machine for servicing requests.
You might want to investigate DRDB which makes a clustered block device (which you could put any file system on top of).
This IBM developer works article gives a pretty good overview of the system. Your question has enough breadth that you could write a book on the topic, so I think all we can do is get you started.
Something like this any use to you looks like you can build your own storage cloud from multiple servers so that upgrades etc will be easier. You can bring down a node while other runs and keeps storage live.