Cathryn
|
Some simple server questionssimple, simple  
We have been advised by our (very nice) IT support providers (who will also sell us the hardware) that we need to replace our second server because one of it's two hard drives is broken. It's six years old (that's old isn't it!).
As an organisation we are often too reactive. Each time a bit of IT goes wrong we buy a new one. This might be a good plan but since no-one has ever considered how we should be making full use of IT and doing some forward planning (an IT strategy ), I am starting to wonder what sort of questions I should be asking including whether or not we should just be renewing our contract with our current provider, nice though they may be.
I am not here asking for free advice, well not quite  just wondering if any of you could start to nudge me in the right direction.
The server could apparently go at any time but my feeling is that even if this happened the world wouldn't end and we do actually have time to consider this a little before rushing off to spend a few thousand of our already frozen budget.
|
tahir
|
What does this server do? How many other servers do you have? What size is your organisation?
|
Cathryn
|
 Great, exactly what I need, thank you.
More please.
|
jema
|
The big question would be whether it really is worth buying rather than renting these days. Renting hives off the maintenance and so it worth thinking about.
|
Treacodactyl
|
As a hard drive will only cost about £30 why do they suggest replacing the whole thing?
What backups do you do, how long would it take to replace the server if it went wrong and would you have everything needed (software, data, licences etc).
|
tahir
|
| Treacodactyl wrote: | | As a hard drive will only cost about £30 why do they suggest replacing the whole thing? |
And is the hard drive part of a RAID?
|
Emyr
|
Depending on what you actually use the servers for, you could replace a few servers with one good server acting as a front end for Amazon's S3 storage service and/or EC2 computing power.
Using these will mean that your costs are more proportional to your actual requirements, and should your requirements suddenly increase, the increase in available processing power and storage can be seamless.
From a green point of view, using shared services such as EC2 and S3 may be cleaner than having your own actual servers.
|
Cathryn
|
Thank you, some of these questions I have already tentatively suggested to my colleagues* but I can ask them with more authority now (in that they don't know the answers either and so shouldn't just take the immediate things offered.) Emyr, I don't understand most of what you've put but I will look them all up and attempt to follow it.
*I have learnt an awful lot more about IT having been on downsizer and I am grateful to you all for this.
|
Emyr
|
Amazon S3 is a storage service, used by (amongst others) Twitpic, the tool which lets twitter upload images and simultaneously announce them on Twitter.
By using S3, Twitpic doesn't have to worry about growing out of their servers and Service Level Agreements mean if Amazon lets them down, they pay less for the service.
EC2 is a similar scheme but for running programs instead of storing files.
|
Calli
|
Cathryn - try search for cloud hosting?
Amazon are the supermarket suppliers 
|
jema
|
Still need to know more about what the servers do?
Cloud based solutions are good for light use.
|
vegplot
|
You'll need to consider your fiscal and budgetary controls. Leasing is good from a taxation point of view (whether is be Amazon or physical hardware) but in some circumstances capital expenditure may be more appropriate.
My preference has been to separate services and data storage. That is have a inexpensive server hosting the OS and then add robust attached storage (direct or cloud) with backup. This allows you to replicate the server and OS for redundancy and keep costs as low as possible.
Also, you need to consider number of users, their location (in or out of the local area network), data throughputs etc.
|
jocorless
|
What software are you running and how much up time do you need?
I've just been working with Amazon S3 over the last month or so and whilst very good we have some outstanding issues that just won't go away and there doesn't seem to be a massive incentive to solve them from Amazon - you have to solve them yourself or through the forums
Which for me as a techy is simply a frustrating challenge but to a non-techy type - it could floor them completely - So be careful - We aren't yet happy to use it as our production storage service
Sounds like what you really need as an organisation is a full review of your systems and what they are really doing - There is lots out there to help productivity and collaboration without spending vast amounts of cash
Only trouble is - finding a suitably qualified consultant who will be honest and not try and sell you something that you don't really need
|
Cathryn
|
You are all brilliant! I am going to extract all these questions and put them to my colleagues to consider. There are other political issues about what the organisation might look like in a month or so time and I am hoping that we can put off any decision until then. Then we can properly address this.
Please can I ask if the statement below is accurate. It uses a RAID 1 configuration (?) because if we could just replace the hard drive at little expense as someone up the thread suggested might be possible, then this might be the best short term option.
If this drive also fails, Office and CAMS applications will not work and remote sessions will not be possible. Also if there is any data on this hard drive there is a high possibility that it will be lost. Unfortunately due to the age of the server (six years old (double the industry standard life expectancy and ROI return on investment period which is typically three years)), XXX would not recommend replacing the hard drive, as past experience with this situation has proved to be costly for past clients, both in time and monetary expense (old drives are expensive to buy as there is limited supply, there is no guarantee that a second hard drive will successfully re-Raid with the first and any issue encountered will cause additional downtime and increase considerably the cost).
|
vegplot
|
| Cathryn wrote: | You are all brilliant! I am going to extract all these questions and put them to my colleagues to consider. There are other political issues about what the organisation might look like in a month or so time and I am hoping that we can put off any decision until then. Then we can properly address this.
Please can I ask if the statement below is accurate. It uses a RAID 1 configuration (?) because if we could just replace the hard drive at little expense as someone up the thread suggested might be possible, then this might be the best short term option.
If this drive also fails, Office and CAMS applications will not work and remote sessions will not be possible. Also if there is any data on this hard drive there is a high possibility that it will be lost. Unfortunately due to the age of the server (six years old (double the industry standard life expectancy and ROI return on investment period which is typically three years)), XXX would not recommend replacing the hard drive, as past experience with this situation has proved to be costly for past clients, both in time and monetary expense (old drives are expensive to buy as there is limited supply, there is no guarantee that a second hard drive will successfully re-Raid with the first and any issue encountered will cause additional downtime and increase considerably the cost). |
I would go with a new system with a minimum of RAID1 on the OS disk rather than replacing the existing hard drive. Adding RAID to an existing older system may not be easy or financially viable given the lost cost of new servers and their better performance. As servers age the hard drive is not the only thing more item which will start to fail. This is a good time to re-evaluate your entire systems.
|
Cathryn
|
Thank you.
|
Barefoot Andrew
|
Drive performance has certainly moved forward in the past few years, but a drive of six years ago won't be that alien technology-wise to what we have now.
Sounds to me like they've had duff experiences in the past in this arena, and are preferring to avoid the scenario again rather that a drive replacement necessarily being the wrong thing to do.
A.
|
