Keith via plug on 27 Apr 2021 13:17:17 -0700


[Date Prev] [Date Next] [Thread Prev] [Thread Next] [Date Index] [Thread Index]

Re: [PLUG] backup lessons from a cloud storage disaster


On 4/27/21 1:12 PM, Rich Freeman via plug wrote:

On Tue, Apr 27, 2021 at 11:08 AM JP Vossen via plug
<plug@lists.phillylinux.org> wrote:
I like, "There is no cloud, there is only someone else's data center."

I agree, because it actually encapsulates the pros and cons of the
situation.  I don't think you can talk about "the cloud" in black and
white terms and I think a lot of people in the FOSS community/etc
don't stop and think about how normal people interact with technology.

The average person keeps all their stuff on a desktop PC with a single
hard drive, and one day it stops working right and they call their IT
friend up and ask them to help them get their photos back.

For THIS average person, the cloud is a MUCH better solution.  Sure,
they're still putting all their eggs in one basket, but it is a
professionally-managed basket vs the PC sitting under their desk
buried in dust without a UPS.

This is a pretty fair statement.  The only thing I was add is that professional doesn't mean what people thinks it means. Professional often implies a superior level of skill or experience but classic definition of meaning you are paid for your services is most appropriate.  Clearly, in this case the hazards that this company succumbed to were a result of not having the experience to consider their own failure points or not having the skill to architecture a more resilient solution.  Ultimately this a 1st order failure by the guy with the other computer (or data center).  Such things are obvious risks.

I don't do residential business but for my strategic partners and others, I always advise to have 2 or 3 physical copies of their data that they can get access to in a reasonable amount of time and then if you want, get a 3 or 4th copy in some cloudy thingy.

I'd defer to Keith and a few other consulting-types here, but I
suspect most small businesses (who aren't paying Keith/etc) are in
similar shape.  For them the cloud is also a better solution than what
they have.

You're 100% right.  A big part of what I end up having to do is push back on the concept and use of cloud.  First explaining that it is more of a marketing term than anything else and second that you are doing nothing more that using resources at a distance. Third, if you do embrace the use of resources at a distance, consider why you aren't allocating your own resources in a remote data center.  That's usually where arguments break down and stakeholder realize that their motivations are not as sound as they thought they were.  Things really go sideways when I (and most people here) can build a solution that is going to make more sense.

That said, I have to admit that even with the expansion of attack surfaces due to the COVID-19 pandemic (i.e. remote work) and supposedly a heightened sensitivity to business continuity, I'm finding that too often organizations are not willing to invest in the resources to protect their operations from data losses and complete system failures.  I continue to not understand that and this is like year 11 for me.

Now, for those on this list who have a RAID and an offsite backup
solution and regular testing and all that good stuff, the cloud may
very well be a step back in a lot of ways.  We're not typical IT users
though.

We PLUG'ers are so elite.  People just don't understand the words coming out of our collective mouths  LOL

Seriously, though, you are right.  The funny thing is I think we all have have these conversation professionally or privately and have they have gone over people heads as to why its important.

I think for most the practical solution is probably some kind of
combination of the cloud plus local storage.  Either it is local
storage backed up by a cloud service (Backblaze or whatever), or a
cloud service (Google Photos/Drive/etc) with local backup (Google
Takeout/etc).  That is pretty convenient to set up and protects you if
the cloud provider goes down, but also protects you if YOU go down.

For people on this list there are a lot of other options that I think
I've talked about.  Personally right now I use a combination of ZFS,
lizardfs, bacula, and duplicity+gpg+AWS.  The first two are primary
storage, the last two are backups.  Duplicity is run by cron daily so
that ensures I always have a cloud backup as a fallback, and it is
encrypted so privacy isn't really a concern here (obviously I have
offsite copies of my key as well - and that key isn't used as anything
else so basically it acts like a second factor in addition to my AWS
credentials).

I've talked about my solution too, which really is the a bit of the same that I use for my clients so I won't burn bandwidth restating it but the most important think you said is that it is a combination.  Cloud / remote storage is fine and local storage is fine.  When you combine the two you get a solution that far more robust that either methodology on its own.  Circling back to the op's article-  the simple point that those clients that had other backup resources in place where fine though this failure can not be amplified enough.

--
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Keith C. Perry, MS E.E.
Managing Member, DAO Technologies LLC
(O) +1.215.525.4165 x2033
(M) +1.215.432.5167
www.daotechnologies.com

___________________________________________________________________________
Philadelphia Linux Users Group         --        http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion  --   http://lists.phillylinux.org/mailman/listinfo/plug