1

There are five types of backup methods for your data.

  • Copy
  • Daily
  • Differential
  • Incremental
  • Normal

You can exercise all or just a few of these methods, whichever is right for your organizational standards or policies, but how many backup locations do you need for your data?

Do you backup locally on your machine, on a DAS, NAS,(locally) and in “The Cloud”. If you utilize a DAS, NAS, how many local backups do you maintain? If you backup to the Cloud, how many locations? How many backups and locations do you use and why this choice?

21 Spice ups
2

Perhaps a better starting point is “What is the aim of the backup?”

Accidental file deletions are handled by shadow copies.

Local DAS or NAS can handle long-term backup retention, depending on the business needs, and rapid local disaster recovery.

Offsite storage can handle disaster recovery when the original building is out of service.

11 Spice ups
3

I was always told it should not be a “backup strategy” for DR, but a “recovery strategy” and they are not the same for every organization. I think the rule of 3 is a pretty good starting point.

8 Spice ups
4

3-2-1 works quite well I think:

  • 3 copies
  • 2 different types of media
  • 1 offsite copy

There’s no single right answer though, budget, value and sensitivity of data, recovery objectives, quantity of data etc. all play a part.

15 Spice ups
5

Luckily, I have six branches. I placed a NAS at each branch for branch local documents.
I backup (Robocopy) each one to the HQ NAS once per week and differential backup (Robocopy) every weekday evening.
I also Robocopy my HQ backup files to one of the branches every weekend.

3 Spice ups
6

We have 3. Shadow, backup to disk, and backup to tape.

1 Spice up
7

I don’t believe there is any such thing as “too many backups”. I use our corporate-defined app for D2D backups and duplicate to tape for storage. I also have a couple external USB disks that I use Robocopy to keep copies of many document folders on for quick recoveries. I also have my preferred (or ‘dream’) backup apps installed (both Unitrends and Veeam) which I use to back up specific servers - this allows me to perform “oops-recovery” the old-fashioned way with the corporate procedure, then I perform the same recovery in a fraction of the time with one of my preferred apps, then send the results to the powers-that-be so they can see real-world examples of how much better we can handle restores & disaster recovery with other solutions.

Hasn’t helped change the corporate mentality on backup apps/procedures yet, but makes me feel good about showing them how life could be much better for us if we were allowed to implement a REAL disaster recovery tool.

3 Spice ups
8

In addition to how many backups, here’s an article I came across on DarkReading where they say “organizations that employ three or more vendors actually lose more data and more money”.

Screen_Shot_2015-01-02_at_11.16.24_AM.png

How many solutions do you typically have running?

1 Spice up
9

JoeWilliams, Rockn, & hutchingsp have all posed great questions and suggestions. Are you trying to create a backup or DR plan?

10

pfft.

What was the root cause of the loss of data or cost of downtime? Can it truly be attributed directly to having 3 or more vendors? Or, what I think is more likely, did the losses occur because the administrators responsible for data restores & disaster recovery were not familiar enough with the procedures to prevent the data/revenue loss?

The infographic for each nation in EMC’s report claims the top-scoring leaders “use a range of modern data protection methods” … that statement sounds to me like the leaders use different vendors for premier protection of various types of data. One size does not fit all, especially when it comes to data protection and business continuity.

11

We have two servers in different locations that we run backups to, capable of spinning up a VM or two if a host goes down. Snapshots every hour and differential nightly offsite. Continuum is what we use, been pretty solid.

12

i have 3-2-1 plan for my file servers and DAG system and veeam replication for exchange and i am stil not satisfied. On Christmas eve, my physical backup server’s RAID card blown and my vm backup sever corrupted… luckly, no one worked after that. i had enough time to repair and configure it.

1 Spice up
13

talking about bad luck!
A christmass to remember

14

Backups are so important in IT that more is better (IMO).

That said, I think 10 full backups a day would be excessive. I’m referring to a full disk backup, full tape backup, duplicated full backup from 1 site to another etc. once or twice a week…it really does cover your behind in the case of a DR

15

Personally I think the best way to look at this is in three separate ways.

  1. You ideally want some form of fast recovery solution (such as virtual replication etc) which can get you out of a hardware failure quickly and with the least possible interruption.

  2. Next is the offline backup where you are using some form of backup solution to take a full backup of your environment. This is obviously very useful in the event of something major going wrong or having to revert back to data that is several days/weeks/months old etc but because of the lengthier restore time option 1 is certainly a recommended first step.

  3. It is really essential to make an offsite copy of the offline backup just in case something serious such as DC’s burning down occurs. Personally, we now use tape/external drive backups very rarely for our clients now and instead offer 2nd DPM Servers to replicate the initial server. This works very well and we find offers much greater flexibility and increased backup frequency over other methods.

In general, for clients who are very conscious about resiliency we recommend;

-Multiple Key Servers across different office locations where possible so that if one server/office is lost the other can take over (Hyper-V Replication works very well for this in our experience.)

-DPM at one of the sites to act as the primary backup solution.

-2nd DPM within our DC to backup the 1st DC. There are a couple of main reasons for the 2nd DPM being at our DC and not at another clients site (although this would still be an option) which are;

If for some reason the client has a major issue which prevents them bringing their backups online, we already have the backups in our DC and so can bring a virtual copy of their infrastructure online from there for them.

If for some reason the client is subject to some form of deliberate sabotage the backup in our DC is external from this and therefore more secure than if it was housed onsite.

This is our way of looking at things but there are obviously many different things that people want to achieve so in general it all comes down to requirements. As long as they are all managed correctly then I can’t really ever see too many backups as being a bad thing unless they were either insecure or hampering performance.

1 Spice up
16

Ummmm… why only 5 backup methods? Why those 5? What does ‘normal’ even mean these days?

17

Our off site backup is sitting beside my bed…

18

We are currently doing 30 day retention of backups at 2 physical ‘hot’ locations and ‘forever’ retention at a third cold storage physical location.

I would eventually like to keep 3 copies of data at 3 physical locations with a ‘forever’ retention, but this can be excessively costly :slight_smile:

Currently we do daily incremental and weekly masters.

I consider DR as protecting and restoring service in a timely manner, dictated by business RTO.

Backups as protecting and restoring data, dictated by business RPO.

We are more concerned with data loss than service loss :slight_smile:

19

As a few have mentioned. Think of this from a recovery stand point. What is your organizations tolerance for loss?. If your server crashes, how long before it can be restored and how much potential data has been lost? You may not have the same answer for all sites, servers, or files.

We gave up on Backup exec, robocopy, etc. and have moved to a Barracuda Backup appliance with cloud storage. The initial backups took multiple days but now take minutes to run.e have incremented our backups to run multiple times a day to reduce the exposure of loss from days to hours for most instances.

The system continues to need to be monitored and quickly resolved when errors occur. There is not a solution that can be installed and only reviewed once a week unless you can tolerate large loss of data.

20

I think hutchungsp nailed it.

But I would advise against a “recovery only” focus. If you don’t get high quality consistent backups then you are not going to have a solid recovery. Backup and Recovery must go together. If you are looking for a good place to start developing a plan, I might suggest this article I wrote a little while ago: " Backup Basics: What do SLO, RPO, RTO, VRO and GRO Mean? ".
Lastly to answer the original question… Yes you can have too many backups. I can relay several horror stories about the wrong backup data set being used for recovery at the wrong time which lead to extra long RPOs (see above). You need multiple copies but not infinite copies.
-George

2 Spice ups