One sad day the trusty server in the corner of the office will go dark. Although you can’t predict when that day will be, you’ll want to be in control when it happens (and plan for it never to happen).
There has been an adopted rule of thumb across the industry of how often a server should be replaced. This currently sits at 3-5 years but isn’t necessarily as cut and dried an answer as it first may appear.
What’s Inside a Server?
To understand how long a server should last, we should firstly look at what makes up a server.
Server components are not, on the whole, mechanical. The CPUs, memory and motherboards are not made of moving parts that can wear out after an extended period of use. If properly serviced, cooled and protected from electrical surges, they can continue to function for a long period of time.
If your server contains Hard Disk Drives (HDDs) these contain spinning discs, so will wear out sooner or later – sooner if you read and write a high quantity of data regularly.
On the other hand, your server may be new enough to have been fitted with Solid State Drives (SSDs). These are non-moving and have a much longer life span.
In either case, drives can be replaced, ideally before they fail, extending the life of the remaining components.
This throws into question the 3-5 year lifespan statement.
When to Consider Replacing a Server?
If we no longer take 3-5 years as the hard and fast rule based on the wear and tear of parts, you’ll rightly be wondering where the figure came from. You don’t want to be retiring a perfectly good piece of (costly) equipment for no reason, plus it’s not the most environmentally friendly approach, if that’s of relevance to your business or personal policies.
There’s a lot more to look into when considering replacing your server than simply how long it’s been up and running.
Warranties typically run for three to five years – correlating to the lifespan figure. When the server is out of warranty the business is liable for all service costs and sourcing of parts. Depending on how new a model your server was when you purchased it, parts may no longer be available, therefore second-hand items may have to be found on sites such as eBay. Coupled with this, finding a maintenance engineer to both diagnose any faults or fix any issues may come with a cost higher than you may have anticipated. They will have almost free rein to charge what they like as they’re providing a rare service pivotal to your business operations.
A basic cost/benefit analysis should give an informed decision as to when it no longer becomes cost-effective to maintain the ageing server and to resign or reassign it before a fault occurs.
The advancement of technology may be the factor that necessitates a refresh. Your server may have been running technology that was a few years old at time of purchase, therefore if you’ve had it for 5 years you could be working with 8 year old technology by this point. 8 years in technology is a long time.
A new model may have more powerful processors, the ability to host more virtual machines and greater memory capacity. On top of this, a new server may have much more advanced power conservation features, enabling lower daily operating costs.
In software terms as well, you may have started running into compatibility issues depending on if you’ve been upgrading your operating systems in advance of the server.
Your financial modelling may influence when to consider replacing a server. If you’re running a leased server it may be more financially viable to upgrade than to commit to an extension. These can be expensive addons to keep an older piece of equipment up and running. Check your terms and if you don’t own the servers it may make more sense to refresh.
Signs That a Server is About to Fail
We absolutely don’t recommend you rely on these signals when deciding when to refresh your server but they’re handy tips to watch out for.
- Server running hot. Reliability is seriously reduced when a server runs hotter than intended. If you hear the fan going constantly or can physically feel the heat when walking by, it’s time to look for the underlying problem. There could be issues with power supply or memory levels and you’ll need to check all components are running as you expect. Don’t forget the simplest cause of all: dust. Is the front intake blocked, the exhaust blocked, has the server recently moved, do parts of it just need a clean?
- Unexpected reboots or random failures. If your server is sporadically under pressure which causes a failure or a reboot that’s nothing out of the ordinary, sometimes it just all gets a bit too much. However if it’s a normal day with normal operations and you’re experiencing crashes, that is cause for concern. Some tips include, taking a look through the event logs to identify any off behaviour, checking for damage to the interior, checking and reseating the memory components and looking for an infection or intrusion via antivirus/anti malware software.
- Computers running slowly. This is one of the most common issues for an IT technician. If it’s just a natural slowing down over time, then this is nothing unusual. However if a sudden slowdown occurs, this is cause for concern. A process may cause a memory leak to eat up system resources, or bad sectors in the hard drive may result in slow data transfer rates and a bottleneck. As a crude indicator, noises coming from the server are also a warning sign.
What Happens if a Server Fails?
We’ve discussed what makes up a server and considerations for replacement, but what really happens when a server fails. There are many different scenarios, you may be a single, small office that operates everything from the one server, or a large business spanning many floors with multiple servers carrying out independent tasks. What server failure means to you will come down to what backup and disaster recovery provision you have in place. Assuming you’re fully protected and can simply spin up a duplicate server then nothing much happens when a server fails. If appropriate, you could extend its life for as long as suits your budgeting plans.
If you don’t have robust and extensive disaster recovery in place then you have much more to consider. Firstly, what data would you lose and what impact would this have on your operations? How long would it take to get you up and running again? Do you know how much in revenue terms you lose for every minute, hour, or day you’re out of action? Can you afford the risk of an unexpected, open-ended bill? If you’re comfortable with your answer to all these questions then a server failure doesn’t sound like a big deal. However, if you don’t know how to answer some of these, or the answer worries you, then you need to have a think about if a server replacement should eke its way up the to do list.
Diagram: The Benefits of Upgrading
What to Replace the Server With?
You have options. You could go for a direct replacement, just more modern all round, improving performance, gaining the associated benefits of a warranty and giving you hopefully the next 3-5 years of hassle-free computing.
Alternatively, you could take the opportunity to make the move to cloud computing. This would mean removing any on-site servers you had, saving on valuable physical space and power and cooling costs. Your servers would then sit in a central datacentre and you would access them via an internet connection. There are huge benefits associated with moving to the cloud, some of which, along with the process of moving to the cloud, can be found here on our dedicated mtech.cloud page. You can also download our infographic on the Top 6 Reasons to Move to the Cloud. At M-Tech we run our own managed cloud service aimed at finding the right cloud solution for you.
Is it time to think about a replacement server, or even a move to the cloud? Talk to us today.