@Bob I have the data on my new box, just need to actually put him into a container. It'll take me 10 minutes once I actually commit to rolling up my sleeves and doing it

Ah

@bwDraco I was stuck at home due to agoraphobia for a few years. I had crippling social anxiety and I'm still arkward around people, online and off. Its just a matter of stopping to think before you do things.

Oh, @allquicatic, did you get the link to that ACI ep?

If I must be critical, you seem to crave/demand attention a fair bit, and are much too implusive with things like this.

@Bob yeah, and watched it and enjoyed it :)

@allquicatic One of the rare few safe ones!

@JourneymanGeek I may be able to adjust this within a matter of hours...

@Bob per wiki, some ground crew got injured when it blew up, which they didn't mention in the ACI

@bwDraco nobody changes anything about themselves -- really changes -- in a few hours ... doesn't happen ... takes time, reflection, experience

@JourneymanGeek I'll be coding an emergency bugfix to the codebase to address this issue.

:P

You know what I mean.

@JourneymanGeek Woof I want attention

@Dog stop barking, Dog!

lol

@allquicatic Ah :( - still, injuries better than fatalities...

Hey OVH's SSL cert for soyoustart.co.uk is broken

@Dog go fetch. Or something.

@bwDraco saying you will fix it is attention seeking behaviour too

@Dog who cares, lol? their enterprise line is where it's at :D in terms of value for your money etc

@allquicatic Well me, because Chrome won't let me view their website

SYS ones have always been better value for money than their enterprise line... the Enterprise line has always been the worst value for money :-/

@allquicatic SYS is decent for compute power, though, isn't it?

Enterprise line was mostly if you wanted a crapton of RAM and/or IP addresses.

@Bob yeah, true... just... urgh... can't stomach the monthly fees for IPs, poor storage offerings and lowish RAM :P

@Dog depends on what you want, doesn't it? because having lots of RAM and IPs assigned to a box has value of its own

Actually, E3-SSD-4 looks nice.

Dangit, E5-Sat-2 only available in Canadia

I'm sure there's tons of use cases for SYS considering that it's usually sold-out

but it's not for me so therefore no one could ever possibly want it /sarc

Will make sure this never happens again. It's quite easy to do in principle but it does mean that more checks will be in place to ensure attention-seeking behavior is weeded out everywhere.

@allquicatic SYS gives you more RAM and a better processor for less money

hi @dog

oh

As for IPs, you get 16, which is more than I need... not sure what the enterprise ones have, it's not listed on this page

sorry I forgot, hi journeyman geek :)

@Dog Enterprise has no monthly fees for IP addresses at all, just the setup fee ($3 one-off per-IP). I think the limit was 256?

Hello

@Bob SYS has no monthly fees either, just setup fee.s

@Dog Only up to 16. Above 16 is's $3/IP/month

More than I need...

True, more than enough for most :P

hi draco ♥

@Bob Actually the fees are the same

Enterprise gives you the first 256 free, SYS the first 16 free.

After that enterprise charges £2.04 a month as well

uhm I actually came here today because I have something I can't figure out in my linux

@Dog whaaaa?

Or maybe I'm just being stupid

> Up to 256 Failover IPs with no monthly fees*

So you get 256 free but no option for more.

But still, waaaay more than I need

@Dog Up to, yea. You still need to pay the $3 setup per IP. And have justification for usage in the request.

Though if you're hosting a lot of SSL sites without that funky modern extension, they could come in handy

@Dog They probably thought you were running a porn site.

I probably was

I don't think OVH prohibits adult content, compared to some other providers.

The issue is behind us. Let's move on.

>_>

No one's talking about it

@Bob not sure if hosting adult content would be an issue for most of us ;p

Bleh, Xeon D 1540 still costs an absurd amount more than the 1520 at OVH :-(

@JourneymanGeek Hm?

Oh

I didn't see the baleeted post

I meant many hosting providers (shared, VPS, dedi) explicitly prohibit adult content in their AUP/ToS

50% off the E5 V2's is an awesome deal. Except none have hybrid storage

who wants to buy a symantec SSL certificate for 700USD?

ohoho !Dog woof you want attention?

That said, with 128GB of RAM for £40 per month... I could probably shove all my VM's into a RAM disk and not have to care about SSDs

20 cores, 256GB of RAM for £86? Are you kidding me? o_0

you can resell it?

@CausingUnderflowsEverywhere You can get a free SSL cert from StartCom for free!

or letsencrypt

:) I got it from letsencrypt

That OVH pricing is literally insane. I want one. Now.

thanks you Ben though :)

Hmm have to choose between Roubaix and Gravelines...

Dude. Gravelines

that name is coolest name.

come play pawn tactics with me, just register that name and I'll say you're the coolest guy online

Bob don't worry you're not a toiler, no one is

.........quiet.. okay I'll leave. take care everyone

Gravelines used to have 1-2ms lower latency from the UK than Roubaix but now it's basically indistinguishable

0.1ms difference average... from VM it's slightly different with Gravelines being a bit more unstable

Lets try from my phone... best out of 3 eh?

Nope, no difference there either.

Hey @Bob/@allquicatic: Can you ping/traceroute proof.ovh.net and gra.proof.ovh.net and see if there's any difference?

@JourneymanGeek Cooler than ponytelecom?

@Dog more Metal I suppose

I'll start working on learning web development again, so you might start seeing me talking about HTML, CSS, and JS.

@Dog Not easily... my current network blocks ICMP

One sec, I'll try from an Azure VM

...which I have to boot up first

Of course, bigger questions will be asked on SO.

@Bob o_0

Weird. OVH's network map suggests London => Roubaix goes via Gravelines, with more bandwidth to Gravelines than Roubaix. Yet their network weathermap shows no connections between Gravelines and Roubaix at all, and far more bandwidth to the latter.

@ThatBrazilianGato Here in chat.

Yeah, I see it now. >_<

(@Bob, please relay this to @dog)

Pinging proof.ovh.net [188.165.12.106] with 32 bytes of data:
Ping statistics for 188.165.12.106:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 345ms, Maximum = 348ms, Average = 345ms

Pinging gra.proof.ovh.net [5.196.90.200] with 32 bytes of data:
Ping statistics for 5.196.90.200:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 328ms, Maximum = 346ms, Average = 339ms

> ping proof.ovh.net
PING proof.ovh.net(proof.ovh.net (2001:41d0:2:876a::1)) 56 data bytes
[...]
--- proof.ovh.net ping statistics ---
20 packets transmitted, 20 received, 0% packet loss, time 19034ms
rtt min/avg/max/mdev = 72.464/72.625/72.939/0.344 ms

> ping gra.proof.ovh.net
PING gra.proof.ovh.net(2001:41d0:a:79c8::1 (2001:41d0:a:79c8::1)) 56 data bytes
[...]
--- gra.proof.ovh.net ping statistics ---
20 packets transmitted, 20 received, 0% packet loss, time 19027ms
rtt min/avg/max/mdev = 71.981/72.117/72.284/0.263 ms

Hmm, interesting.

This is from Linode in the Newark, NJ datacenter.

Gra being faster but also more inconsistent, same as here

That's via LTE

@Dog well, consistency is... questionable. See: LTE.

There's a difference of about 0.5 seconds in favor of the Gravelines datacenter.

(@Bob, please relay this to @Dog)

Azure VM still starting

@Bob My LTE is very consistent, that said, empty channel, high band, and Fing sends like 100 pings

Ooh it's @ThatBrazilianGato again. Thanks

@Dog Yea, it's a tad noisy here apparently.

That's effin weird, those results you're both getting

From everywhere I can test from, Gra is ~0.5ms quicker to UK and ~1ms slower to EU

btw, bwDraco posted a couple via a Linode server if you want to look

+/-10ms is unexpected

O_O

Gra is also less heavily loaded right now, so you'd expect less variation

I'm testing from Azure now via RDP

The Win10 command line is... uh... typing a character at a time.

@Dog Again? I have no idea what you're talking about... :P

Looks like the typewriter animation on PowerPoint...

Roubaix is far better connected than Gra, so why the hell is Gra managing to be faster and slower at the same time

@Bob Oh lol, that :-/

Looks like for non-Western countries Gra might be taking quite different routes back, even though it's literally one hop and right next to Rbx

Pinging 8.8.8.8 with 32 bytes of data:

Ping statistics for 8.8.8.8:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 1ms, Maximum = 2ms, Average = 1ms


Pinging proof.ovh.net [188.165.12.106] with 32 bytes of data:

Ping statistics for 188.165.12.106:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 275ms, Maximum = 276ms, Average = 275ms


Pinging gra.proof.ovh.net [5.196.90.200] with 32 bytes of data:

That's from Azure, Australia East.

Ffs, I was just settled on Rbx now you guys got me considering Gra again

I'm like Rbx is bigger, older, better connected, and I've used it before... and it's only 0.5ms difference

But Gra seems better for everyone else except me :-/

Lemme grab some traceroutes.

Hm. First four five six seven eight nine hops timing out. That's not a good sign.

I'm seeing stars past the ninth hop.

> traceroute proof.ovh.net
traceroute to proof.ovh.net (188.165.12.106), 30 hops max, 60 byte packets
 1  router1-nac.linode.com (207.99.1.13)  1.001 ms  0.975 ms  0.967 ms
 2  173.255.239.0 (173.255.239.0)  0.938 ms  0.933 ms  0.931 ms
 3  nyiix.nyc.ny.us (198.32.160.77)  1.611 ms 173.255.239.9 (173.255.239.9)  0.869 ms  0.889 ms
 4  * nyiix.nyc.ny.us (198.32.160.77)  1.592 ms *
 5  * be100-5.nwk-5-a9.nj.us (192.99.146.96)  1.968 ms *
 6  be100-1298.ldn-5-a9.uk.eu (192.99.146.132)  69.821 ms be100-5.nwk-5-a9.nj.us (192.99.146.96)  2.436 ms  2.430 ms

Argh, why does everything OVH say contradict itself

If you'd excuse me, I have some work to do.

Yea ok tracert is horribly broken on Azure.

Weird.

If ping works tracert should work too...

Unless it only allows certain ICMP responses?

Possibly

Tracert relies on ICMP error messages, rather than echo replies

> "Roubaix Valley" - the birthplace of OVH - now has more than 100,000 servers spread across no less than 5 datacentres. The entire dedicated hosting range is available; SP, HG, EG, and MG.

> The Gravelines datacentre opened in 2013 and is the biggest datacentre in Europe. When running at full capacity, it will host a record 350-500,000 servers. You have the option of hosting your SP and FS range servers here.

Gravelines = newer, bigger, less connectivity, but they're building a second one, whaat?

@Bob lol

Thanks for trying :-P

Why can't anything ever just be simple -_-

Oh, the half price deal is :"* Uniquement avec l'utilisation du code promo "DEDOVH2016". Tarif valable les trois premiers mois de location"

Because y'know, fuck English.

Right back to the beginning then

@Dog Maybe I should look into a quick AWS instance :\

@Bob Meh, fuggedaboutit

feedback.azure.com/forums/217313-networking/suggestions/…

Maybe I should just bite the bullet and switch to OVH

Sure it costs 20% more but then I get double the RAM plus two 480GB SSDs

superuser.com/review/low-quality-posts/565987

@Ramhound handled

"Woman dies in ambulance crash" Ironic

@Dog this is unrelated, but your chat line reminded me of this... "On August 17, 1957, Richie Ashburn of the Philadelphia Phillies hit spectator Alice Roth with a foul ball, breaking her nose. As Alice was being carried off the field on a stretcher, Ashburn hit her with another foul ball, breaking a bone in her knee."

o.O

He must really hate her.

0_0

Lol

> Luckily, my shoe wasn't on the ground long enough to be glued to the floor, but now, the entire bottle of superglue is stuck to my shoe

Full disclosure. I didn't meant any disrespect by me pointing out the answer was a single paragraph, was a little annoyed, they pinged me asking if it was better after making it even longer though. I then apologized for that misunderstanding. I then clicked on the link only to be pointed to a support.com website (which has nothing to do with HP).

Feel oblgiated to do something with about a support.com scam

(Hp.sxpport.com) isn't actually hp support clearly

Meh

@Ramhound disabled people do not want to be treated any different than any other person, me thinks you covered that :-)

Ok.

Didn't mock him.

Anyways.

@Ramhound felt like trolling. Don't feed em ;p

erf

I should have used a mask. (Just ground down a PVC pipe to fit another pipe with a dremel. Used a vacuum to clean up as I ground, but my throat is all scratchy)

I ought to be wiser when making decisions, giving thought to the effects and alternatives before taking action. This takes self-control and I intend to implement changes to this effect as soon as tomorrow.

This is what I truly admire about you and is what separates the wheat from the chaff.

Thank you for being so patient with me. I've already made minor but visible improvements that would not have been possible without your guidance.

@JourneymanGeek, thank you.

(had input issues because Firefox for Android is so slow and my phone is low on RAM 🙁)

linuxcontainers.org/lxd/news

@allquicatic Hm?

@Bob new lxd release

a few nice bugfixes

also that release apparently enables OpenVPN within a container without any tweaking or hacks

@allquicatic Huh, useful. Going for OpenVZ-parity?

@Bob that has to be at least a long-term goal of LXD

Today the roomie nosed the portable air conditioner extake vent. Waited for me to turn on the air corn. Then glanced back as he walked a little further cause it took a moment to kick in. Then sat with his arse to the warm air coming out ;p

@JourneymanGeek lol

O_O

> As of vzctl 4.0, OpenVZ can work with unpatched Linux 3.x kernels, with a reduced feature set.

> This new release focuses on merging OpenVZ and Virtuozzo source codebase, replacing our own hypervisor by KVM one.

openvz.livejournal.com/42793.html

@allquicatic Looks like OpenVZ also uses LXC now...

And has for four years o.O

> Ability to manage containers and VMs with libvirt and virt-manager or virsh via a single driver for containers and virtual machines.

@Bob github.com/ariya/phantomjs/issues/14458

github.com/annulen/webkit/issues/117 github.com/ariya/phantomjs/pull/14386

TL;DR: work is in progress to move to a very "new" WebKit codebase (enabling a large amount of support for new web platform features) that will require you to build and link against a "system" Qt5 and QtWebKit, but right now headless support may not work and you can't disable web security

@allquicatic Ooh, nice

so you cna keep up to date from now?

@allquicatic ...not nice

the inability to disable web security kills our effort for now, until they implement the disable switch :P

@allquicatic I can't even remember why web security was an issue.

Oh. Was it the loading of script from localhost?

Pretty sure I could just embed the bot script straight on the page instead.

I had high hopes for jBrowserDriver, but it apparently depends on a hacked WebKit that's built into Java 8 and later... and the version of WebKit underlying that is dependent on Oracle / OpenJDK's maintainers, and they don't seem super eager to use the latest webkit

I did a bit of poking around and I think they were still using webkit from 2013 in the latest Java 9 early access

@allquicatic I'm more concerned about the headless-ness. I'm absolutely certain I can hack around the web security issues.

If they're what I remember :P

@Bob I haven't actually tried it, but that open issue on annulen/webkit doesn't inspire confidence

If you can't get headless, might as well run Fx or Cr, yea?

annulen/webkit is a much newer version of webkit for QtWebKit than is available in the upstream Qt sources, so there's no real point in recompiling PJS using upstream QtWebKit, even if it's headless

upstream qt stopped updating WebKit full stop

they moved to QWebPage or something, which uses Chromium, but of course is not fucking headless

because anything that touches Chromium has never been headless

lol

I wonder how much work it would be to make Fx headless.

Of course, no one bothers, cause Xvfb.

wait, what the fuck is this? chromium.googlesource.com/chromium/src.git/+/lkgr/headless/…

lol

chromium.googlesource.com/chromium/src.git/+log/lkgr/headless

@allquicatic Design doc from Feb 2016. Very new, probably not stable.

Also makes me wonder how 'headless' it is, if it uses GL?

@Bob well, tries to use GL -- it might fall back to Angle or something, which IIRC is software rendering

probably faster to render using a heavily optimized software OpenGL than traditional rasterization

Interesting that someone is trying though.

llvmpipe is pretty freaking fast

@allquicatic The fallback is osmesa.

and that thing is bottlenecked by the wonky complicated X11 design

osmesa? :S

But that it supports GL at all in headless mode sounds weird for me. I'm admittedly not very familiar with how X interacts with the GPU...

@allquicatic mesa3d.org/osmesa.html

> Mesa's off-screen interface is used for rendering into user-allocated memory without any sort of window system or operating system dependencies. That is, the GL_FRONT colorbuffer is actually a buffer in main memory, rather than a window on your display.

Oh wait, that's not the fallback.

@Bob hardware GL in headless mode makes no sense; while you can technically use OpenGL without X in KMS with open source graphics drivers, most of the time if you're doing that you have an X server

That's just how they do GL in headless mode. Interesting.

@Bob nice -- and if they're using the llvmpipe Gallium3d backend for the heavy-lifting, that's an LLVM optimized OpenGL software renderer

you can play ~2014 native Linux OpenGL games at 15-30 fps at 1024x768 on a Core i7-4790K with llvmpipe.. it's pretty damn impressive

groups.google.com/a/chromium.org/forum/#!topic/chromium-discuss/…

@allquicatic bugzilla.mozilla.org/show_bug.cgi?id=446591

> Now that rendering is heavily based on Cairo, there's the opportunity to use its "image surfaces" API (in addition to XLib, Quartz or Win32).

Interesting.

github.com/laurentj/slimerjs/issues/80

You +1'd that :P

Meh. Local challenger has an abysmal selection of headsets :(

Man, this competition is crappy

Challenger being the big, non great large computer chain.

@JourneymanGeek Does Lazada have a physical store?

> phantomjs: error while loading shared libraries: libicui18n.so.52: cannot open shared object file: No such file or directory

my old phantomjs doesn't have the libs it needs on Ubuntu 16.04

damnit...

Welcome to DLL-hell, Linux-style.

symlinking doesn't work; they genuinely broke the ABI

phantomjs: symbol lookup error: phantomjs: undefined symbol: UCNV_FROM_U_CALLBACK_ESCAPE_52

> _52

come on, do you really have to break the API that often?

@Bob they're like eBay or qoo10. They run an online storefront for others.

Ah.

This is the sort of thing I'd buy at the mall that closed down.

lol

CH?

!!info

...

!!listcommands

*crickets*

pjs and nodejs are running and he's in here

@allquicatic want me to take a look?

@allquicatic same setup? (at work, cbf typing on phone for steam)

@Bob same, except ubuntu 16.04

I copied over the icu*.52 libs from Ubuntu 14.04 and phantomjs runs

@allquicatic You doing anything on there right now? No screen instances...

*checks server time* juuuust in case :P

[INFO  - 2016-08-16T06:52:04.158Z] GhostDriver - Main - running on port 4444
[INFO  - 2016-08-16T06:52:04.406Z] Session [asdfasdf] - page.settings - {"XSSAuditingEnabled":false,"javascriptCanCloseWindows":true,"javascriptCanOpenWindows":true,"javascriptEnabled":true,"loadImages":true,"localToRemoteUrlAccessEnabled":false,"userAgent":"Mozilla/5.0 (Unknown; Linux x86_64) AppleWebKit/538.1 (KHTML, like Gecko) PhantomJS/2.1.0 Safari/538.1","webSecurityEnabled":true}
[INFO  - 2016-08-16T06:52:04.406Z] Session [asdfasdf] - page.customHeaders:  - {}

that's from debug.sh

!!info

hmmmmm

it's not showing up in bot overflow either

@allquicatic mind if I try?

not at all; killall nodejs && killall phantomjs

not an incredibly recent nodejs version but it should be enough

[ERROR - 2016-08-16T06:55:25.329Z] GhostDriver - main.fail - {"line":85,"sourceURL":"phantomjs://code/main.js","stack":"global code@phantomjs://code/main.js:85:56"}

  phantomjs://platform/console++.js:263 in error

interesting

can you kill -9 my phantomjs please

@allquicatic pretty sure errors are bad :P

trying to exit back to phys and it's hanging

ty

ahhh whoops, killall on nodejs failed so the && did nothing

wait, that may have been cause of your error

yea

works now

I just ran ntpdate time.windows.com on the host

(don't judge, I just memorized the server)

lol

I already checked against time.is

wat do

gimme a sec

nodesource latest stable nodejs?

You are now in a REPL with the remote page. Have fun!
> bot.adapter.out.add('I have just been restarted! This happens daily automatically, or when my owner restarts me. Ready for commands.');
Attempting to run "bot.adapter.out.add('I have just been restarted! This happens daily automatically, or when my owner restarts me. Ready for commands.');"
! Error Can't find variable: bot

that's not a good sign