Tech&ME said:
I think its best not to talk in this topic.(For ME atleast, since everyone is talking based on there DESKTOP experiences and DIGIT Linux distros)
Before replying to this topic I would like to request forum members to atleast tell us.
1. How many of you run real SERVERS (Windows or Linux based) in your real life ?
2. What Critical Servers are and what type of Servers are we talking about here ?
Let me do the honors of answering that.
1. Me, Our team Manages 30-35 Servers ( We do Active monitoring of these servers as they host Various Server Components of the Application We support).
2. Criticality? Hmm.. Well.. A downtime of >5 Minutes of these servers without prior information is considered to be a Serious Business Problem.
Ok, this what we handle.but there are other servers which we dont have access, totalling about 10,000 Servers.
Now Some servers is so critical that they have 2 - 3 backup machines. Why? Because our client is one of the top 3 Financial Assets Management Company, One of the Biggest Bankers in the world. They are so paranoid anout security that we are not allowed to work on our personal desktops, we have to do remote session into their Machines located elsewhere and work. They are so strict that I have signed an agreement that I will not reveal their name anywhere. And couple of guys got sacked, coz they downloaded sometihing off the e-mail.
Enough said about criticallity and Confidentiality.
The Server of Choice for all of em is without qustion Unix. There is nothing which can remotley matchup the scalablity and strength of Unix machines.
The unix servers manage all the critical and vital softwares of the bank.
And no this is not my opinion. Its by people who knows systems like back of their hand. And its for a company where Systems HAVE to be running all the time as they have production servers in 9 Locations. So in a day there is atleast 1 Site is having its business hours.
And my dear friend Windows Machines dont see past 2 weeks uptime. And unix machines see an uptime of 6-8 Months.. where a reboot is did after some patch upgrade or for some maintanence work. OS never crashed. But Some apps with run with kernel ring crashes. So they can be upped only with a full reboot (Also because some issues take onger to resolve without a rebootl. So far we have encountered Two such instances in Unix. And that wasnt a Unix issue.. it was a hardware issue. Interestingly Unix system ran pretty crippled mode with one bank of Rams Failing totally.. (Low performance. no crash)
Oh yeah.. we do have Linux machines too...In fact some of our servers are Linux. Even they are quite good. Uptimes of 4-6 months are quite common...
Now Any questions?