Amazon Power Failure knocks out QRZ for 12 hours

The Cloud is sharing with all.

The "Cloud" is just a voluntary spy system to get your personal data put on the internet.

Be careful what you share with others...

 

How did we know that the cloud we were operating on was a Cumulonimbus.

 

Yep, and that was a cirrus mistake on our part.

 

Some of us would like more cloudy days.
And All QRZ All The Time.

 

Never trust cloud computing with your data. We have always been told it is bullet proof. Famous last words. I am however, amazed that there was no backup plan. Your description seems to indicate (and I can't believe this), that there was no emergency power or U/G data lines. So much for the much touted "Cloud".

This should be a lesson to all.


K2WH

 

We are all electric also. Five solar panels 16 = 12V batteries, 3 inverters, two generators. Neighbors and PUD were down for 54 hours and we did not notice until all the light pollution returned.
But I, too, did miss QRZ. GLAD you are back up and running.

 

Great you are back, but it does beg the obvious question...

Do you really want to put all of your eggs in a cloud system? (not specifically QRZ, but it seems everytime you look around, there is somebody else trying to sell smoke & mirrors by putting stuff in a cloud... cell phone music, kindle books, etc)

There really is something to say for local storage of ALL your data (as you guys used to bring things back on line).

Not buying all the h y p e! Long live the Seagates & Hitachi hard drive mfgs.

Getting more grumpy every day but making multiple backups!

Tim W5FN

 

This was not a single thunderstorm. This was like being inside a weak tornado for a couple of hours. Commerce in my part of Virginia is at a crawl. It is difficult to explain the weather phenomena that took place last night. It was freakish.

 

I agree 100%

Hardware fails and software crashes, this is inevitable, but to bring down multiple environments because of a simple data center power failure is indefensible. In a HA (High Availability) environment with the proper engineering and implementation, clients should never, never be exposed to any infrastructure problems no matter how severe. It will be interesting to see what the ISP has to say about root cause, reports of cascading failures and what DR strategy is in place to protect clients.

73
Tom

 

I was lost!
I had to go & annoy the "Ginger Beers" on EEVBlog,instead,but it wasn't the same!

73,VK6ZGO

 

Yup. It really was.

BTW, for everyone on the thread - the "cloud" doesn't have to be all bad, and every one of you is using "cloud" instances today, know/like it or not.

For instance, here on the 'zed, you log in and can see what threads you've repied to or not. That's "cloud" - it is data you use that you don't store. Your callsign lookup is essentially a "clouded" version of the otherwise publicly available FCC license data.

A larger enterprise will likely rely on a number of public, private and hybrid clouds, and they may be on-premise, off-site or co-located. Email and data storage are often "clouded" into some variant. The main thing that differentiates a cloud from a "managed solution" is that the cloud is purchased by some unit of consumption...like per bandwidth used, or Tbs of storage used. If the unit is time, then the units tend to be small, like minutes or hours. (if the time unit is a year or more, you essentially have a lease and not a cloud. )

Interestingly, the Bell phone system before the breakup was essentially a "cloud" service - you rented the phones and paid generally by how much you used the phone. You could walk away at any time, and someone else owned and worked to keep up the infrastructure. That's cloud in a nutshell.

Today, many things beyond data storage are now "cloudied" - compute power, bandwidth, email, VoIP, chat, etc. These all tend to be unique/separate clouds, so while a company may embrace could services for its needs, it will NOT have all its eggs in one basket. Yes, an event like this storm might knock out access for a few hours, but it will be to a single aspect, like data retrieval. Phones, internet and other aspects will still work. IF, on the other hand, they kept those functions in-house and their own datacenter went down, they may actually lose ALL those important functions at once - phones, data, etc. Believe it or not, when a company moves to cloud solutions it is generally mitigating the risk of a single point of failure, not increasing it.

There are risks and rewards; there are also risks and rewards for localized data too. In the long run, I suspect we'll continue to see some "clouding" of more stuff but also believe there's a natural limit, as some things don't lend themselves to utility (by the drip) pricing.

What I'm merely doing is pointing out that the cloud itself isn't 100% bad and it is fairly widespread.

 

The internet does have many, many nodes. Amazon's datacenter is not the internet, it is connected to it. Amazon's datacenter failed, the internet did not.

The cloud paradigm is not contrary to that, it is a complement to it - unless, like Amazon, that particular cloud fails either because of lack of redundancy or the redundancy is completely overwhelmed.

To say "keep everything on your own drive" is the least safe and robust way to do it. Sometimes for security reasons it is the best choice overall, but there are severe tradeoffs when doing so.

 

Yeah, the storms did pack quite the punch!

 

I didn't see it coming...as I didn't know QRZ was offline (since I was too). But I do find it entertaining.

 

My UPS works very well at home. Never have a problem with losing any data when the power goes out. Has Amazon heard of these devices??