Strokes may seem harmless at first, but they're often deadly when left untreated. Early diagnosis is key to reversing the symptoms and achieving full recovery.
The same is true for problems with your company's server. Did you know more than 90 percent of all server issues can be caught and corrected before they become serious? Watch out for the "STROKE" warning signs, so you can get proper treatment for your server right away. Healthy systems keep downtime to a bare minimum.
"S": Suspicious sounds
When the power goes out, one of the most noticeable things is the sudden quiet. We become accustomed to certain sounds, such as the air conditioning and fans, the icemaker in the fridge or the ticking of a clock. When those sounds stop, we notice the change.
If you notice this same phenomenon with your server - new sounds, changed sounds or the sudden absence of sounds - consider it the equivalent of a fire alarm going off. It could be a fan that stopped or a hard drive about to bite the dust. Ignore these problems at your own peril, because mechanical problems can shut down your server completely. Avoid this with spare parts and prompt attention from a qualified professional.
People are also reading…
"T": Temperature change
Servers are pretty particular about their environment. Dust and heat are bad. If you notice your server room or even your server are warmer than usual, pay attention. Server rooms must be cooled and ventilated, and servers themselves require a number of internal fans to keep comfortable. A room temperature of 70 degrees or less is ideal. Keep a thermometer in the room as an extra precaution.
If any part of the cooling system isn't working, your server can overheat. Fried eggs are yummy; fried circuit boards are just a waste, not to mention a fire hazard. Yes, it's rare, but they really can catch fire.
A quick side note: We've seen more than one server room with sprinklers installed overhead. Fire is bad enough; add water to the mix, and your systems really will turn to toast. If you want a fire suppression system for your server room, look at gas-based alternatives.
"R": Red lights
Servers have plenty of flashing lights, but flashing red lights on peripherals such as your uninterruptible power supply (UPS) can be a sign of problems. UPS batteries can go bad, which means you have no power protection. Red lights on routers, hubs and switches often indicate communication failures. Red lights on backup devices may signal a failed backup process, which means you can't recover data if you experience other serious problems.
Any time you think you notice a change in all those blinking lights, even if it's just a gut instinct, it's a good idea to have everything checked out. Your server and systems are made up of many interconnected parts, and any one failure can cause a ripple effect that becomes steadily more serious.
"O": Outdated software
As your server ages, so does your software. A major key to smooth-running systems is regular application of patches and updates. This includes your operating system, as well as software applications and security systems, such as your anti-virus program.
If these aren't kept up-to-date, or if the updates stop working for any reason, your system becomes vulnerable to hackers and security breaches. The consequences? Someone could steal all your data or corrupt your server to the point where it needs to be completely erased, rebuilt and restored. You do have backups to restore, right?
"K": Komplaints (or Known Issues)
Whenever your staff start to accumulate a list of intermittent and occasional computer problems, it can be a sign of something more serious going on. They are noticing changes in how your systems operate; unless you've deliberately made updates, these may be unplanned and unwanted problems and a sign of worse to come.
Keep a central log of computer and system problems. This will allow you to watch for trends and patterns. Repeat problems or issues affecting a number of people need to be watched closely.
"E": Errors
Most server problems start small and escalate, but your server keeps track of what's going on in the error logs. Regular checks and analysis of the messages can alert you to issues long before they become serious. Many of these error messages are cryptic, but early diagnosis can make the difference between a simple repair and several days downtime while you wait for new parts or even a new server.
Need a doctor?
You wouldn't treat a stroke on your own; unless you enjoy downtime, you shouldn't tinker with your server either. These warning signs can alert you to issues, but you'll need to bring in experienced server techs to perform whatever surgeries are needed to correct the underlying problems.
Round-the-clock care
Want your techs to show up on your doorstep before you even have time to call? New tools allow remote monitoring of your systems, like round-the-clock nursing care watching over your business 24/7. Your techs will be alerted to problems as they arise, often before you even realize there's a problem. Many systems make it possible to fix issues remotely as well, which saves on travel time and gains you faster response.
A proactive approach like this nearly eliminates downtime, and you won't have to stress yourself by watching out for server "STROKE" symptoms every day. Keep your systems healthy, and you can rest easy.
Wendy Gauntt is president of CIO Services, LLC, a technology consulting company that specializes in small business solutions. Visit her Web site at www.cioservicesllc.com for more information, ideas, and free resources. Copyright © 2008 CIO Services LLC
