You think it will never happen to you. You assume your hard
drive, server, network or perhaps your entire data centre is safe from
unexpected outages. But when IT disaster strikes, it comes without warning, and
can be in some bizarre forms.
Here are some of the weirder IT outages from over the
Flood? Sure. Fire? Undoubtedly. But squirrels? Oh, yes.
Attacks from our furry friends can drive data centre operators nuts. A squirrel
took down a Yahoo data centre in 2010 after chewing through electrical cable.
In 1987, a squirrel shut down the power system feeding the National Association
of Securities Dealers’ automated quotation service for 82 minutes, stopping 20
million shares from being traded. The financial system survived, but the squirrel
didn’t. Squirrels are among the top source of US power outages. The American
Public Power Association even indexes them.
landing for BA’s power equipment
In 2017, over the May bank holiday weekend, British Airways
was forced to ground thousands of flights after what it described as a power
surge knocked its data centre offline, taking down baggage, ticketing and
check-in systems. 75,000 passengers were affected. CEO Alex Cruz said: that the
back-up system “did not work at that particular point in time." The cost?
takes out Rackspace
A Dallas/Fort Worth facility owned by data centre company
Rackspace lost power to its chillers in 2007 after a trucker ran into a utility
pole, knocking it into a transformer. Rackspace switched to a second utility
provider to power its data centre’s chillers, but they failed to restart,
causing it to take some servers offline in the rising heat. The incident took
out services from SaaS provider 37Signals.
error causes VM terror
A fat-fingered employee at cloud services provider Joyent
simultaneously rebooted all virtual servers in its US-East-1 data centre in
2014, causing outages for users. As all the virtual servers started booting
back up, the load on the system slowed everything down. It took up to an hour
for everything to get up and running again. The whole incident left the company
investigating how a single operator could cause an outage of that magnitude
without any failsafe mechanisms kicking in.
getting...err...warm in here
In June 2018, Microsoft's Azure suffered a service outage
in northern Europe which impacted customers for at least five hours. And the
cause? Microsoft's North Europe data centre is in Dublin, and the temperatures
that day had soared to the heady heights of 18C... yes, just about t-shirt
According to Microsoft an increase in outside air
temperature, meant that control systems kicked in to keep the temperatures
within the ideal range. However, this lead to an unexpected increase in
humidity levels that caused hardware component failures in multiple network
devices and hard disks, and a lot of very frustrated customers.
second, let me fix this
In 2012, the International Earth Rotation and Reference and
System Service added a leap second to the Earth's time to compensate for its
uneven rotation. Unfortunately, it caused the Linux kernel to lock up, throwing
many computer systems into turmoil. Reddit and LinkedIn were hit, as was the
Amadeus Altea system that provides booking and reservation services for many
airlines. The system was disrupted for over two hours, throwing airlines
including Qantas in Australia into disarray, cancelling hundreds of flights.
Rather than taking any chances, investing in an up to date
disaster recovery solution can help protect your precious data from furry
fiends and fat fingers. If you haven't invested in proper disaster protection
already, best look at it right away. After all, with random events like these,
who knows what might crop up in your computer systems?