Backing Up To The Cloud
Backups are the one thing you HAVE to get right. Backing up to the cloud can help.
Tapes where pretty much the only way of backing up for years. Slow and cumbersome, these were then replaced for many by hard drives – though pretty much used the same way.
Taking last night’s backup off site (home with you), leaving it in the car over night so that the moisture could get at it, forgetting to bring it back in the morning if you had remembered to take it in the house, dropping it on the pavement outside your house, Baby Joe trying to eat it …… just some of the of the many backup stories I’ve heard over the years.
Backing up to the Cloud -all that seems like ancient history – but apart from the physicality’s of tapes and hard drives, best practices remains the same.
Whether you’re platform is Linux, OSX, Windows or something else – there are two stages to managing successfully backups – planning and administration.
Planning Your Backups
Planning may seem the simplest step – identify the servers, workstations and files you want to back up and then which product to use. You’re ready to go right? Whoa – hold your horses, there are a few things to consider.
Impact Analysis (IA)
Not all servers, users or files are equal. Analysing which parts of the business are most critical, and which servers and data are most critical to those is an important first step to ensuring your Backup & Disaster Recover Plan is going to work when needed. Of course you may just have one server or just a couple of laptops but many businesses rely on multiple devices all of which provide different functions to the business.
Recovery Point Objective (RPO)
Just how much data can you afford to lose? Is taking daily backups enough? Can you afford one recovery point a day, maybe you need two? One at lunchtime and one in the evening. Is your data so critical you need to take incremental backups every hour or even more often? But for which systems? As with Impact Analysis not all data is equal.
Data Retention Policy (DRP)
How long do you need to keep the data for? How far back do you need to be able to go? A week (a little short), a month (more common), seven years due to industry regulations?
Recovery Time Objectives (RPO)
So you’re backups are taken twice a day but one of your servers has died completely. How quickly do you need everything restoring? In the old days backing up to tape meant restoring would take as long as the tape drive manual would say – maybe in Kbs – adding on the time it took Bob to go home because he forgot last nights tapes.
Backing up to the cloud means your internet connection speed is critical to the recovery time. Do you need a faster line? Will your Backup Cloud Provider be able to send you a rider with a copy of your data on a hard drive? Can your provider spin up a virtual machine in their datacentre with an image of your server? Should you have an onsite NAS with 3 days data for quick restores to augment the Cloud Backup?
None of the above is difficult – and if you’ve been doing backups for years you’ll be familiar with the concepts. The biggest impact of backing up to the cloud is how long will it take to get your data back (and of course how long will your initial backup take). Hence a local NAS or a full DR service where a provider can spin up VMs with your servers may be something to consider.
Picking a Provider
Once you’ve identified your requirements you can go looking for a suitable provider. With all the points above documented it will be much easier to ensure you find a provider that can meet your criteria. There are so many out there.
Ask what benefits your provider give you that a local tape or hard drive backup cant and take advantage of them.
Some providers offer a managed service – where they will install the backup agents, create the jobs, check for errors and do the restores for you whilst some are completely hands off. What suits you best?
Administering Your Backups.
Just because you’re backing up to the cloud does not mean you can just fire and forget. Good Practice for tape and hard drive backups still apply when backing up to the cloud. Making the assumption that everything is OK because it’s somebody elses service is a dereliction of duty.
Know Your Backup State
Ensure that you receive regular (daily) reports of any backup jobs that have been created, whether this is by email or by logging into your backup portal. Don’t just check that they’ve run – check for errors and drill down into them ….. and make sure you fix them even if the current failures are not important. One of the gravest mistakes I’ve observed (multiple times) is someone digging into an error – deciding the error is not critical and then ignoring the error flag on the backup job the next day and the next because they assumed it was the same error. Oh dear oh dear. You only make that mistake once – but it’s better not to make it.
With GDPR just around the corner – everyone needs to take security seriously. Ensure you receive a copy of the provider’s GDPR policy statement, general security policies and confirmation of ISO 27001. You really don’t want to trust your data to a provider who can’t provide you with these. It may also be important to identify where your data is stored – i.e. in which state or country depending on the industry you work in.
Good Practice, that’s what will enable you to sleep peacefully at night, and at the heart of administering backups is test restores. Test restores prove that your backups are good and that they can be restored in a timely manner. Weekly test restores should be good enough for most people. Testing restoring different files, from different servers and workstations at different times is the best way to go about this. And when the files have restored (to a different test folder of course), open them. Don’t assume just because they have found there way back to your hard drive that they’re actually intact.
Scary Backup Stories
Whilst it’s quite a long way from Halloween, here are some stories I can personally verify that may keep you up at night.
Many many many years ago when I was just starting out in the world of IT, we had a customer using the venerable ICL System 25 (at least I remember it as that model). One day they had a complete loss of data and brought out their their backup tape (singular) to restore but lo-and behold the data had gone AWOL. Every night for 5 years they had dutifully put the tape in the system and let the backup run. Back then the system didn’t verify the data after a backup. They hadn’t bothered to do a test restore in all that time and when they needed it – it was empty. The theory at the time was that the tape had been completely de-magnetized. Early computer systems weren’t very clever and it hadn’t noticed the error and has just continued to written the data to oblivion.
Not too long ago this one, but long enough ago to be backing up to 3.5” floppy disks. One evening a little fire broke out in the warehouse. That little fire grew into a big fire and soon consumed the office where the account computers could be found.
Resting on top of one of them was the box of backup floppies (no hometime for these little babies). Needless to say they didn’t withstand the inferno well and two weeks later the business tanked.
I’ll Get Around To It.
Bringing us almost up-to-date, imagine not the most competent IT Manager being told by the company yours’s truly use to work for that their backups hadn’t worked. Notifications sent after 3 days – then a week, then two weeks – then a month – then two month …… this went on for about six months, including many telephone calls. After about six months their system was struck with an encryption virus that merrily made it’s way around the network encrypting any piece of data it could find – which was most of it.
Queue a panicked IT manager asking for whatever we could restore – which was nothing. This was followed by a call from an irate CEO threatening all sorts. After the email trail of all the failure notifications had been forwarded all went quiet. The IT Manager did not last long.
Backups are your friend. Treat them with respect and due diligence and they’ll sve your bacon time after time, and the company you work for.
Backing up to the cloud is a great way to take the hassle out of backups, ensuring consistent off-site copies which with the right service provider could spin up all your servers in the cloud if you have a complete loss of building.