So, what exactly goes into an effective disaster recovery plan?
Well, it depends. Keep in mind,
there's no one-size-fits-all for disaster recovery plans.
The mechanisms chosen and procedures put in place will
depend a lot on the specifics of your organization and environment.
But we can't start by covering the three types of measures in more detail.
We'll also go over some examples to help give you an idea of what to think about.
A good way to understand what to plan for is to perform a risk assessment.
This involves taking a long hard look at
the operations and characteristics of your teams.
A risk assessment allows you to prioritize certain aspects of
the organizations that are more at risk if there's an unforeseen event.
Risk assessment can involve brainstorming hypothetical scenarios and analyzing
these events to understand how they'd impact your organization and operations.
When you look into preventive measures,
pay attention to systems that lack redundancy.
If it's something critical to permit operations,
they should probably have a redundant spare, just in case.
Make sure you have a sound backup and a recovery system,
along with a good strategy in place.
Ideally, you should have
regular but automated backups to backup systems located both on site and off site.
It's also critical that you have
data recovery procedures clearly documented and kept up-to-date.
Data recovery following a disaster is hopefully something you will rarely do.
It's important to make sure that these procedures are
updated since systems change and evolve throughout their lifetime.
Redundancies shouldn't be limited only to systems.
Anything critical to operations should be made redundant whenever possible.
This includes power delivery or supply,
communications systems, data links, and hardware.
Think through the impacts that a disaster affecting
each of these aspects would have on your day-to-day operations.
What would happen to your network if the building lost power?
Can you use continue to work at
the fiber-optic data line if the building gets damaged by a nearby construction work?
You call router for the office just burst into flames.
Okay. The last one is a little far-fetched,
but I think you get where I'm going.
Another super important preventive measure that should be
evaluated and verified is operational documentation.
Make sure that every important operational procedure is documented and accessible.
This includes things like setting up and configuring critical systems and infrastructure.
Any steps or specific configuration details that are needed to restore
100 percent functionality to core systems and services should be documented in detail.
It's also important that this documentation is kept up-to-date.
An effective way to do this is periodically
verify that the steps documented actually work.
You don't want to find yourself in a situation where
you need to reconfigure assistant following incident,
only to discover that the documentation is wrong.
When looking at detection measures,
you'll want to make sure you have a comprehensive system in place that can quickly
detect and alert you to service outages or abnormal environmental conditions.
If up-time and availability is important for your organization,
you'll likely have two internet connections;
a primary and a secondary.
You want to monitor the connection status of both of these links.
Ideally, they should be configured to automatically fail over if one goes down.
But you'll still want to be alerted when this happens,
so you can investigate why the link went
down and work on getting it back as soon as possible.
The way to think about designing detection measures is to evaluate
what's most critical to the day-to-day functioning of the organization.
What infrastructure services and access are absolutely vital?
Those are the ones you should be closely monitoring.
And it's not just outages or complete failures you should be watching for.
Of course, you want to monitor for those,
but you also want to monitor for
conditions that indicate that a problem is likely to occur.
If we can avoid a catastrophic failure by
being alerted to an overheating server before it fails,
that would be much better, wouldn't it?
So, you want to monitor conditions of service and infrastructure equipment.
Things like temperatures, CPU load and network load for
a service monitoring for error rates and requests per
second will give you insight into the performance of the system.
You should investigate any unusual spikes or unexpected increases.
These early warning systems allow you to head off
disaster before it brings operations to a halt.
And of course, you absolutely must test these systems.
Simulate the conditions your monitoring systems are designed to catch.
Make sure the detection thresholds actually fire the alerts like they're supposed to.
This testing goes beyond just the monitoring systems too.
You'll also want to test your reactions and responses to these alerts.
If you're monitoring systems reliably trigger alerts but everyone just ignores them,
they aren't super useful, are they?
You want to conduct regular disaster test to make sure systems are
functioning and that your procedures to handle them are also up to the task.
Corrective or recovery measures include actions that are taken to restore
normal operations and to recover from an incident or outage.
This includes taps like restoring a corrupted database from
a backup or rebuilding and re-configuring a server.
Your disaster plan should include reference or
links to documentation for these types of tasks.
Anything and everything that would be required to
restore normal operations following some disaster.
This is where the steps for restoring various systems and data from backups should be.
The disaster recovery doc doesn't need to contain the details of the operations.
Links and references are sufficient.
But it's important to be prepared for a situation where
typical documentation acts methods are unavailable.
Let's imagine you keep all your operational documentation
in a wiki on a dedicated server.
What happens when that server suffers an outage?
It's important that critical documentation is accessible
if an emergency scenario or disaster strikes.
In the next lesson, we'll cover what to do after a failure. See you there.
Google
