Erlang Getting Too-Big-To-Fail Process Flag

From Joe Armstrong comes news that Erlang will soon feature a new process flag for those processes that just really need memory, or else: "Too big to fail processes behave like regular processes until they get too big and memory congestion occurs. If a memory allocation error is triggered when a too_big_to_fail process needs more memory, then a random smaller process is killed, and the system reattempts memory allocation for the too_big_to_fail process. An interesting situation can occur if the too big to fail process has killed all other processes and still cannot get enough memory. In this case the node running the process tries to memory steal from other nodes." Read below for your FREE logged-in-reader's-eye view of the special rot-39 version!

Sebz Wbr Nezfgebat pbzrf arjf gung Reynat jvyy fbba srngher n arj cebprff synt sbe gubfr cebprffrf gung whfg ernyyl arrq zrzbel, be ryfr: "Gbb ovt gb snvy cebprffrf orunir yvxr erthyne cebprffrf hagvy gurl trg gbb ovt naq zrzbel pbatrfgvba bpphef. Vs n zrzbel nyybpngvba reebe vf gevttrerq jura n gbb_ovt_gb_snvy cebprff arrqf zber zrzbel, gura n enaqbz fznyyre cebprff vf xvyyrq, naq gur flfgrz ernggrzcgf zrzbel nyybpngvba sbe gur gbb_ovt_gb_snvy cebprff. Na vagrerfgvat fvghngvba pna bpphe vs gur gbb ovt gb snvy cebprff unf xvyyrq nyy bgure cebprffrf naq fgvyy pnaabg trg rabhtu zrzbel. Va guvf pnfr gur abqr ehaavat gur cebprff gevrf gb zrzbel fgrny sebz bgure abqrf."

Damn it

[phantomfive]

Slashdot keeps making me log in to see ROT13. This April Fool's thing has gone too far. I want my normal ROT13 fare!

Re:

[bryan1945]

I remember when the jokes were clever.

Decryption

[DarkHorseman]

It doesn't work on this story...

ROT-39, Just in the nick o' time!

[girlinatrainingbra]

Hey, they rolled out the special ROT-26 and bonus ROT-39 editions for the last minute before midnight Zulu-time April Fool's Day Edition!
.
Yipee-kai-yay, mother-fokkers! Srsly, what took you so long? What? The day's over alreadY? ... :>(

Great!

[The MAZZTer]

I assume developers will have to pay extra for use of this feature, ensuring ll the large companies get to slap it on all their processes and no one else does.

I mean, ensuring the feature is used responsibly and not abused.

Re:Great!

[wierd_w]

So, this gets to be used by all those bank and loan institutions who use trade bots to manipulate the global stock market?

Question: what happens if 2 such processes re running concurrently on the same node, and actively try to outperform the other by proactivally allocating all available free memory?

Re:

[wierd_w]

But that would imply that a too_big_to_fail process can indeed fail! Allowing the contention to kill the first process, instead of jumping instantly to the allocator of last resort breaks the model!

Re:

[aix tom]

It's kinda like the "things that can not break" in mechanics, I guess.

There the difference between "things that can break" and "things that never ever possibly can break" is: Every time a thing that never ever possibly can break breaks, it's impossible to get at and repair.

Re:

[gl4ss]

I can see Apple or Microsoft offering "priority" CPU, GPU, and Memory access for "premium" apps.

they already do. see windows rt, windows phone and iOS. approved api access(including killing other processes) for "partners" and whoever the fuck is in fashion this week at the middle manager level.

ahhh, now I can divide by zero

[swschrad]

and I don't have to have error checking on my compilers any more. time to get back into programming!

If this had been the only joke...

[istartedi]

If this had been the only joke today, I think it would have worked.

Re:

[phantomfive]

The point of Slashdot on April fool's day is to collect the jokes from around the web. Not much original content here ever.

Hope /. learnt its lesson.

[140Mandak262Jamuna]

None of the discussions today broke 3 digits on comments. Try something like this again, I am not coming back. Stupid rot13 trick. Whoever came up with the idea of running it this long should be fired summarily.

Re:Hope /. learnt its lesson.

[ultranova]

Stupid rot13 trick. Whoever came up with the idea of running it this long should be fired summarily.

Workplaces where people are fired sumarily for mistakes are ones where no one ever does anything except cover their ass and search for any other employer. The quality and amount of work that gets done reflects this.

Re:

[Pieroxy]

It can always be worse. Always.

Re:

[TeknoHog]

Workplaces where people are fired sumarily for mistakes

Ding! You're out!

Re:

[LWATCDR]

true but April Fools is done. It is no longer clever or funny. So Slashdot needs to move on.

Re:

[Ksevio]

That's not actually true...The dinosaur [slashdot.org], Commodore 64 [slashdot.org], and Linus Win9 [slashdot.org] articles all got into triple digits, not to mention a couple plain text ones.

Though I admit I avoided slashdot today because it didn't have real articles and was too much work to read the fake ones.

Resistance certain.

[EricTheGreen]

Angry developers, seeing this as little more than a back-door raid on their precious computing resources, are banding together in newly-formed TCP Parties, determined to resist meddling compiler makers and nebulous language specification agencies answerable to no one....

Vigil - A More Elegant Solution

[Anonymous Coward]

Improper utilization of limited resources is a symptom of incorrect code.

Vigil already has a much more elegant solution to the general problem of incorrect code:
https://github.com/munificent/vigil

Rather than punish other programs, it simply deletes the offending code.

All Vigil programs are guaranteed to run without error, eventually.

Inspired by OOM killer?

[sheepweevil]

This sounds a lot like the Linux OOM killer. You can set the oomadj value to protect certain processes against the OOM killer, making it almost the same as these too-big-to-fail processes. I guess the only difference is stealing memory from other nodes.