This is my first time in this forum so go easy on me please
I am running a dedicated apache server with Plesk. I want to understand whats going on (much) more than I do but I can’t seem to find tutorials or anything close to giving me answers. I am trying to stretch my learning beyond my needs so I have a (sort of), buffer.
When setting up client space/web space/email addresses/client contact details etc, this is stored in a MySQL db and some of it is stored in a dir called ‘psa’. But where should all that data be stored if I wasn’t using Plesk but, instead, were to input it (to where), using my own scripts which already capture ‘all’ the necessary data.
Is there a server management tool that would work on multiple servers connected through a RAID5 array?
If anyone is willing to help I would appreciate it greatly.
OK, guys thanks for that. Maybe I don’t need RAID at all then?
What I need is to have a number of web servers which host clients websites. Because I am familiar with Plesk, I would prefer to be able to use it across these servers.
I then need to ensure that if one of those webservers goes down; that another can step into its place to maintain uptime.
I then need a NAS or DAS database server and I guess another beside it which is a ‘live slave’ ; again, ready to jump into live action with an accurate db, if the primary NAS/DAS server fails.
Is that a straightforward requirement? I do have a server management company in mind for this but I need to know a bit about what I am thinking before getting lost in a world of jargon or, worse, ignorance; and the consequent extra costs.
[list][]Maintaining its own database of your configuration, so that it can present it to you through the web interface without having to analyze the system to figure it out
[]Running commands on your system and editing configuration files of various services[/list]
The domain is set up for serving a webpage by editing configuration files included by Apache, user accounts are created on the system, DNS entries are added by creating/modifying zone files, etc. If you want to do it all yourself, you have to look at the documentation for the services that make up the system – apache for web, mysql for database, BIND for DNS, postfix or qmail for e-mail, etc.
A RAID array is a way of using multiple physical hard drives as if they are one. It is not some kind of networking equipment to connect computers.
Just one thing… answer 2 - I knew that RAID was as you describe (even if I can’t visiualise how it would work) but, I have been told that Plesk can only work on a single server environment. So, in a nutshell my question is; is there a server management tool like Plesk which will manage the different servers, connected through a RAID array.
Why do you add the “connected through a RAID array” part? You can’t connect things through hard drives. Assuming there was a control panel to manage the configuration of multiple servers, what would your hard disk setup have to do with it?
Indeed! RAID is a Redundant Array of IDe drives. It’s purpose is to mirror data stored on the primary RAID disk on at least one other disk, not to provide a “cloud” (load sharing) environment. In other words, RAID should be INVISIBLE to Plesk, Apache, yadda-yadda.
You just need a load balancer that can detect when one of the servers behind it dies. You can use nginx which is free and does just that. Set it up to proxy for two servers, and set the weight of one server very high so that it’s always used. If that server goes down, it’ll start proxying for your backup server instead.
SitePoint uses nginx to serve most of the images on their site.
I don’t know of a “Plesk for multiple servers”. It shouldn’t be too much for you to set up the client twice, once on each server. The MySQL databases can be mirrored automatically through replication.
–
All that aside, are you sure you need such a heavy and expensive setup for your clients? Are they doing some really intensive stuff? If they’re just the average client websites, then you might consider using a VPS account or two instead of all these physical servers. With a bit of scripting you could even use a single Amazon EC2 instance with Amazon CloudWatch to spin up a new instance when the first goes down and assign it the dead instance’s IP address and virtual hard disk (from Amazon Elastic Block Storage). You essentially have a hot backup that you don’t pay for because it’s not running until the main server dies.
The entire OS, and all the service configurations, can be on the EBS virtual disk which allows you to move it from instance to instance and clone it at will for backups.
