Time and Server Conditions Can Endanger Your Data
A few weeks ago, our IT Consultants replaced a Windows Server at a customer site. The 11-year-old server still ran Windows Server 2003. It had become so slow that it choked parts of the network, and its backups kept failing.
To put it bluntly, the server was near death. And could take the customer’s data with it!
Servers Have Lifespans
Like most things, servers have a lifespan. How long is that lifespan? It depends on the role it serves in your network and on the physical environment in which it operates. A moderate- to high-use server (such as a database server or a domain controller) has a recommended lifespan of 3-5 years.
Many factors influence a server's lifespan, and any one of those factors can lead to a shorter life. Some of the biggest factors involved are:
- Age of Operating System. Is the software getting older? Has IT maintained it well?
- Security Status. Does the server have sufficient security in place? Is it kept up to date?
- Server Environment. Is the server housed in a temperature-controlled environment? Are servers stacked properly for good heat dissipation? If the environment is more industrial, does the equipment throw dust or metal shavings into the air?
Despite these factors, many small/mid-size businesses are pushing their servers to 8, even 10 years of operation these days. As those businesses see it, if it still works, why should we pay to replace it?
Well, for one thing, the server could crash at any moment.
Aging servers in warm, dusty environments will crash more easily & frequently. Could it happen to you? Yes. When? If you're relying on an old server, sooner than you want it to.
What Happens When a Server Fails?
There comes a time when the probability of server failure rises from ‘Possible’ to ‘Certain.’ Then you have serious infrastructure problems. Not just one kind of problem either—server failures can take many forms, affecting several parts of your business.
These are the most common server failure types. Along with what usually happens as a result.
- Hard Drive Crash. Data is lost; restore from backup needed. Server could be down for several days while replacing drive & recovering data (assuming your backup is viable).
- Internet Connection/Routing Failure. A routing error disrupts the server's connection, so it appears down. The server may still run, but can't provide its service to the network or Internet. Can cause duplicate copies of data, which require manual fixes.
- Web Service/DNS Failure. Service could be overloaded, too busy to process, or has received an error from DNS. Server needs reconfiguration. Could be down for hours before you notice unless a monitoring system is in use.
- CPU Meltdown. Server crashes hard and won't reboot. You might be able to swap the hard drive to another server—but you're still down in the meantime.
- Memory. One or more memory chips can fail and affect the server. Results are slower application presentation and can result in a bound server requiring a hard shutdown. This can damage data and affect applications like financials which have not committed records at the time of the shutdown.
- Hacking. The server experiences a hack or malware attack. Operations are disrupted. Server applications are damaged. If ransomware was involved, the server may need a complete rebuild. Hopefully you have viable backups.
- Server Hardware Death. The server fails slowly over time and eventually dies. Often caused by overheating. Parts or full server replacement needed.
In every one of these failures, you're left with serious consequences. Data loss. Disruption of business (and sales). IT support & replacement costs.
Server crashes can end up as one of the most expensive business failures. Compounded by the fact that they are very often preventable, just by replacing critical servers every 3-5 years on schedule.
How to Tell if Your Servers Need Replacing
In order to avoid a full-on server crash, it's best to replace servers before they fail. That's a whole new issue of course—how do you know when it's time to replace a server?
Here are a few red flags based on our own experience replacing customers’ failed servers.
- Have you used the server for 3 years or more?
- Have you experienced more than one notable problem with it in the past 6 months?
- When was its last upgrade? More than 6 months ago?
- Is its operating system nearing "end of life"?
- Has the server’s performance noticeably declined over the past 6 months?
- Do users report frequent network slowdowns when trying to work?
For more on what affects the life of a server, visit these links:
Ars Technica: How Long Do You Keep Your Servers?
Lifecycle Information for Microsoft Server Products
Pay Attention to Server Lifespan, and Minimize the Chance of Disaster
If your server is at risk, don't wait for it to fail. Upgrade or replace it before disaster strikes. It’s less costly in the long run. The expense is also predictable, since it’s planned for.
Two final pieces of advice: One, invest in server monitoring or employ Managed Services. This will enable your support team to have insight into hardware issues and allow them to make repairs before a crash in most cases.
Two, implement Cloud Backups. On-site backups are great, but not always reliable and in the event of a localized disaster can be rendered unusable. Cloud backups give you an offsite option without the cost of high end backup solutions.
PlanetMagpie rotates servers in our datacenter every 3 years. Not only does this avoid crashes from old servers, it preserves our hardware warranties. Just in case.
Next month we’ll discuss when to use on-premise servers, and when it makes sense to move to the cloud. Don’t miss it!
Is it time to replace your old servers? Please send your questions to firstname.lastname@example.org.