SoylentNews is people
SoylentNews
The New York Stock Exchange unexpectedly shut down trading in all of its listed stocks Wednesday morning.
The exchange said on Twitter: "The issue we are experiencing is an internal technical issue and is not the result of a cyber breach."
A trader on the floor of the exchange in lower Manhattan, who spoke on the condition of anonymity, said that after the suspension began, traders were told that the problem was related to updated software that was rolled out before markets opened Wednesday.
Trading was resumed around 3:00 PM after having been down since 11:30 AM.
Original Submission
This discussion has been archived.
No new comments can be posted.
New York Stock Exchange Shut Down Temporarily Due to "Internal Technical Issue"
|
Log In/Create an Account
| Top
| 34 comments
| Search Discussion
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
You humans are all alike.
(Score: 2) by mcgrew on Thursday July 09 2015, @03:34PM
I took a college class around the turn of the century, and the instructor was in charge of the Illinois Secretary of State system and gave the class a tour through the mainframe. It ran Unix, of course, had four separate CPUs for redundancy; if one went down, the other three automatically took over its functions seamlessly. They have two natural gas backup generators for if the power goes out.
He claimed that they had achieved 100% uptime since they had installed it, and said that 100% uptime was critical, since the Illinois State Police relied on it for their squad cars; if it went down, none of the squad cars could do anything with the car's computers and it would be dangerous for the cops and maybe even the public.
I would think the stock exchange should learn from my old instructor. 99% won't do.
mcgrewbooks.com mcgrew.info nooze.org
(Score: 3, Interesting) by MrGuy on Thursday July 09 2015, @03:54PM
There is no such thing as 100% uptime.
Even "five nines" uptime is incredibly hard and expensive. And, sorry, having more than one machine and a generator hardly qualifies as 5 9's-level reliability.
What happens if the building that all these CPU's live in suffers a connectivity failure? What if this happens again [wikipedia.org]? What happens if ComEd's natural gas system goes out at the same time the power does, so there's no gas to burn? Or, if you have a store of gas, what happens when the gas runs out? What if the storage the CPU's are attached to fails? Or gets corrupted because all the machines are up but one is writing garbage to the shared repository? Or if the shared repository locks up? Or, if each node has its own local storage, what happens when they get out of sync? Where's your defense against those things?
How does the system get upgraded without downtime? Sure, there are ways to do it, but just having more than one node isn't sufficient to guarantee that.
Also, designing a system that results in near-total failure if one key system (ANY one key system) goes down is begging for trouble. Your system is brittle, and while this does focus a lot of attention and effort on protecting your single point of failure, it means you have one.
Sure, the state of the art has advanced a lot in the last 15 years, but none of these problems was unknown at the time.
If your instructor claimed he had a 100% uptime system, then I think your old instructor didn't know what he was talking about. And if he thought his system was 100% bulletproof, he shouldn't have been in charge of a system. Reliability is a journey, not a destination. You're never done, and you're never perfect.
(Score: 1) by xav on Thursday July 09 2015, @05:31PM
> There is no such thing as 100% uptime.
> If your instructor claimed he had a 100% uptime system (...)
I think you are mixing up HA with uptime. Saying that you have _achieved_ "100% uptime" (as quoted by the poster) doesn't mean you have a 100% HA system. My first understanding is that you just have been lucky enough until now.
Beside, after replacing uptime with HA, everything you said is correct.
(Score: 2) by mechanicjay on Thursday July 09 2015, @06:29PM
I'd argue about state of the art with regards to clustering and availability. VMS, once a prominent player in financial exchanges and whatnot, had a lot of this stuff figured out decades ago. To the point where there are documented cases of VMS clusters having cluster updates of DECADES.
We also had some pretty darn reliable Netware clusters back where I used to work. These things were expensive and proprietary and fell out of favor in the marketplace for a variety of reasons. When messing around with some of the linux heartbeat and HA stuff, I'm astounded that it's so shitty when the problem was basically solved elsewhere a long time ago.
My VMS box beat up your Windows box.
(Score: 2) by Kromagv0 on Thursday July 09 2015, @04:02PM
Sounds like some of the gear we have in my company's building. There is the corporate IT manages stuff (I work for a huge faceless fortune 50 company so we have corporate IT and then the people I work with) that seems to manage 95% reliability and then there are the systems I work with. Some of them have been running continuously for ~13 years since they moved into this building save for patches and the necessary reboots, but even then only one machine goes down at a time and the rest of the system takes over. Multiple redundancy is key, then again the work I do is actually overseen by all sorts of regulators who actually care depending on who the customer.
T-Shirts and bumper stickers [zazzle.com] to offend someone
(Score: 2) by tangomargarine on Thursday July 09 2015, @06:18PM
I was waiting for the part of the story when, during the tour, somebody frobbed a Molly switch in a particularly redundancy-defeating way.
And then the 100% was no more
"Is that really true?" "I just spent the last hour telling you to think for yourself! Didn't you hear anything I said?"