Hey there. I gave a talk at PGConfNYC '14 earlier this year about our solution to HA using open source solutions. The animations and stuff don't really work too well on Slideshare, but it should point you in the right direction:
pgSQL doesn't have anything built in for fencing, failing-over, etc. by default, but by using stuff like Pacemaker, you can get the job done with a little elbow grease.
Since this needs to be enabled in the kernel at compile time, it really depends on what your favorite distribution chooses to include.
If it does end up being in the Big Two (Debian-flavored, RH-flavored) it really doesn't make sense for servers. If it was enabled by default this seems like something that you would need to do some sort of 'handshake' between to prevent someone from just mounting a USB device willy-nilly as they see fit. Complete speculation on my part though.
It absolutely does make sense for servers. One frequent pain point with virtualized systems is that (at least in a VMware vCenter DRS-enabled cluster), you never know what physical host a specific VM will be on at any moment. This is a problem for software that requires a USB license dongle (many FlexLM-managed software have this requirement).
Anyway, in this situation, you would have one physical, bare-metal linux "usb dongle server", which then shares out its USB devices to one or more other linux VMs. After doing this, VMs can migrate between physical hosts without losing access to their license dongle.
There are purpose-built physical USB->IP devices out out there now, but they're quite expensive. This new functionality would allow admins to emulate this functionality for a fraction of the cost of a purpose-built device.
Ah, that's a use case I didn't even think of. Makes sense in that situation then. I rarely deal with things that require USB license dongles, but I can see that as being a pain this may alleviate.
Thunderbolt hotplug is supposed to be handled by the firmware. But Apple decided to implement thunderbolt at the operating system level. The firmare only initializes thunderbolt devices that are present at boot time. This driver enables hotplug of thunderbolt of non-chained thunderbolt devices on Apple systems with a cactus ridge controller. This first patch adds the Kconfig file as well the parts of the driver which talk directly to the hardware (that is pci device setup, interrupt handling and RX/TX ring management).
Finally. Been waiting a while for this. The fact that Apple designed the hotplugging at the OS-level infuriated me for the longest time.
It seems like a legitimate tradeoff; implementing hotplugging at the OS-level gives you DR capability and far more control over the process. I'm not saying they made the right decision, just saying that it appears to be a legitimate one. Obviously the disadvantage is that hot-plugging doesn't work in other operating systems when run on Apple hardware, but obviously that's not Apple's focus or interest.
Intel was willing to certify Apple's devices, so clearly the specification doesn't actually require hot-plugging to be implemented in firmware. So I wonder what the author bases that assertion on.
> Obviously the disadvantage is that hot-plugging doesn't work in other operating systems when run on Apple hardware, but obviously that's not Apple's focus or interest.
Ah dang! I have an official Thunderbolt to Ethernet adapter on my MacBook Pro that won't be detected in windows 8.1 running in bootcamp* unless I reboot.
And now I know why.
> The fact that Apple designed the hotplugging at the OS-level infuriated me for the longest time.
On the flip side, if there's an issue with the protocol, it can be easily fixed with an OS update. Consider that in contrast with the recent USB security hoopla.
You do realize that Thunderbolt has direct memory access right? It doesn't even need drivers to do malicious actions. It can dump and modify memory at will.
In fact all the USB security hoopla also affects Thunderbolt devices, which can also be reflashed without physically modification.
Thunderbolt is even less secure than firewire, which at least required some support from the OS to dump memory.
USB is an iron fortress compared to those kinds of attacks.
Disclaimer: I've never used Lyft and the following is purely hearsay:
In the area I live in (where we've got Uber(x), Hailo, and Lyft) there was word that cars that had the pink mustache were being targeted by taxi drivers for harassment, so I think they pulled them off for safety.
So back in the day, IANA was the sole controlling body for IP address allocations. When the internet began to grow, they redelegated certain blocks (with the largest being /8 in CIDR notation) to organizations in geographical communities. This is called the RIR (Regional Internet Registry) system. In order to get IP addresses from a certain RIR, you have to be a business or customer within that region. The region of North America and some of Latin America is under the American Registry for Internet Numbers, also known as ARIN.
Which all has exactly zero to do with what is meant here by "non-US IPv4 address space", as it in no way implies that the systems that you assign those addresses to are themselves located in those geographic areas, let alone that their users speak a particular language and similar crap that people have overloaded the meaning of IP addresses with. It's a division for administrative purposes, nothing more, and as you might notice, it doesn't even have country granularity.
(Also, the question was rhetorical - I know how that idiocy happened, and I expect that most people here do.)
What about using alternative frameworks for HTTP servers for Python? Pointing at only Bottle and going "hey! Python is slower!" is like pointing at WeBRICK and going "Hey! Ruby is slower!" when you haven't even touched Passenger or Unicorn.
Does anybody use PyPy, Jython or IronPython for a production web service? I'm seriously wondering.
I've had the misfortune of having had to use Jython, but just to read Python pickles from a Java service (don't ask). Even that simple use case wasn't fun (the last stable release was also over a year ago).
PyPy looks great for small or new projects, but there is this disclaimer: "PyPy has alpha/beta-level support for the CPython C API, however, as of 2.1 release this feature is not yet complete. Many libraries will require a bit of effort to work..." - this has caused PyPy to be incompatible with every large Python project I've ever worked on.
And well, personal (dis)taste in .NET aside, IronPython hasn't had a release in over a year either...
Where are these projects being used in production?
I'm even more skeptical now that I see a company tried moving to PyPy, and actually hired a core PyPy developer (Alex Gaynor - http://pypy.org/people.html) to make the move, and it still didn't work out.
As someone who used to work in the web hosting industry, this article completely ignores the technical side of things of handling _THAT_ much attention.
I've seen startups be hosted on shared hosting plans, get a boat load of traffic, the site gets suspended, and they are dumbfounded as to why it happened!
If you spend the amount of money to launch a start up, get a beefy server, load balance it with at least Varnish, and do some load testing before going on mainstream media. Nothing is more embarrassing as when your site tanks after being advertised on National television.
Well, my advice would kind of be the opposite: don't expect a TV appearance to send your traffic skyrocketing. TV is just not a traffic generator. One major reason: they probably won't give out your URL. Even if you get them to agree to drop the URL in, they will probably forget, or not have time, or they'll get it wrong. Even if they do, it's not going to get you much type-in traffic.
Unless you're going on something like Colbert, just don't expect a boost. And even if you do go on Colbert, most of your traffic will still probably come from /r/colbert.
Very true so far for us. Our iPad game got a really nice review on the Canadian show 'Reviews on the Run', and that day I think we had only a handful of Canadian downloads (and American). Nothing out of the norm heh.
Fair - It's helpful to ask around, if you can, re: what to expect ahead of time. I've seen Techcrunch posts cause a bigger spike than a Fox national TV appearance (3000-4000 visits, 1-2000 simultaneous) depending on the time of day, and how much the segment is about your company vs about a trend or a topic that your company is commenting on
http://www.9bis.net/kitty/?page=Welcome&zone=en