| 00:46.45 | *** join/#wowace Seerah (~umsin@2601:3c2:101:320:64f9:56d6:4219:6d7f) |
| 02:42.23 | *** part/#wowace Seerah (~umsin@2601:3c2:101:320:64f9:56d6:4219:6d7f) |
| 04:07.08 | *** join/#wowace Marcy (~textual@162-223-221-68.cpe.ngl.ninja) |
| 04:46.15 | *** join/#wowace ShadniX (dagger@p5B3D20F3.dip0.t-ipconnect.de) |
| 07:20.13 | *** join/#wowace EthanCentaurai (~EthanCent@host-78-147-99-38.as13285.net) |
| 10:00.43 | *** join/#wowace Exec (~me@2001:41d0:401:3100::6d6a) |
| 10:08.14 | zarnivoop | serves you wrong |
| 10:11.15 | *** join/#wowace ls- (~ls@180.183.93.171) |
| 10:30.38 | quiescens | o.o |
| 10:50.00 | *** join/#wowace stolenlegacy (~stolenleg@213-47-49-27.cable.dynamic.surfer.at) |
| 10:50.01 | *** join/#wowace stolenlegacy (~stolenleg@unaffiliated/stolenlegacy) |
| 11:37.03 | *** join/#wowace starfox (~wig0r@193-81-178-119.adsl.highway.telekom.at) |
| 11:49.26 | *** join/#wowace Justwait (~Justwait@546B1128.cm-12-4a.dynamic.ziggo.nl) |
| 11:57.29 | *** join/#wowace starfox (~wig0r@193-81-178-119.adsl.highway.telekom.at) |
| 12:35.37 | *** join/#wowace ZaB|SHC| (~Zaphod_Be@ipbcc04e31.dynamic.kabel-deutschland.de) |
| 12:39.46 | *** join/#wowace ZaB|SHC| (~Zaphod_Be@ipbcc04e31.dynamic.kabel-deutschland.de) |
| 13:25.26 | *** join/#wowace Justwait (~Justwait@546B1128.cm-12-4a.dynamic.ziggo.nl) |
| 13:59.50 | znf | well, fucking reat |
| 14:00.11 | *** join/#wowace Ressy (~Ressy@CPEa84e3f42f293-CMa84e3f42f290.cpe.net.cable.rogers.com) |
| 14:00.11 | *** join/#wowace Ressy (~Ressy@WoWUIDev/WoWAce/ARL/Ressy) |
| 14:00.13 | znf | our .ro TLD administrator (RoTLD) just announced they're introducing an anual maintenance fee |
| 14:00.33 | znf | but nobody (not even the official resellers) know enough details |
| 14:00.48 | znf | previously, our .ro domains were registered for an "undetermined" time period |
| 14:01.18 | znf | so something like 39⬠per .ro domain, for an undetermined period |
| 14:02.08 | znf | now this was a recipe for disaster - there are thousands of abandoned .ro domains and you can't get them back |
| 14:02.36 | wink | it was a one-time purchase? wow |
| 14:02.40 | wink | "purchase" |
| 14:02.42 | znf | yeah, completely retarded |
| 14:03.05 | wink | well, then it can only get better? :P |
| 14:03.53 | znf | anyway, they announced that they will introduce an annual payment of 6â¬/year from 1st of March 2018, so you will pay 30⬠up-front for a period of 5 years |
| 14:04.11 | znf | which is still annoying, but whatever |
| 14:04.38 | znf | they also announced that domains registered from 2012+ on, will "expire" at the end of their 5 year period |
| 14:04.54 | znf | so if you registered a domain on 1.1.2013, it will expire on 1.1.2018 |
| 14:05.15 | znf | if they're older than 2012, then you will need to pay up from... 1st of July 2017 |
| 14:05.24 | znf | ...they're announcing this on 7th of August 2017 |
| 14:05.28 | znf | notice the issue? |
| 14:05.40 | wink | rofl |
| 14:05.44 | znf | and that's about all the information |
| 14:05.48 | wink | still not too bad |
| 14:05.55 | wink | unless you're a domain grabber |
| 14:05.57 | znf | previously RoTLD actually handled domain registration, too |
| 14:06.15 | znf | oh, don't get me wrong, I love the idea of being able to get some old domains which have been abandoned for years |
| 14:06.19 | znf | but it's so confusing |
| 14:06.20 | znf | because: |
| 14:06.30 | znf | 1. will RoTLD still handle domain registration? |
| 14:07.02 | quiescens | i nominate znf to handle the domain registration |
| 14:07.04 | znf | 2. What is the price offered for domain resellers? If RoTLD does it at 6â¬/year, resellers need to get it cheaper, obviously, right? |
| 14:07.54 | znf | 3. if 1 is false, what happens to the domains registered directly at RoTLD? Will they need to be transfered? |
| 14:09.50 | znf | 4. I'm a 3rd hand .ro reseller (because I don't register that many so I don't have "official" API access), what happens to my domains that were registered trough another reseller? |
| 14:10.49 | nevcairiel | so stop being an impatient bastard and give them time to (a) collect questions, (b) answer them =p |
| 14:11.00 | quiescens | o.o |
| 14:11.13 | quiescens | gives nevcairiel a cookie |
| 14:11.27 | znf | nevcairiel, notice my issue with "1st of July 2017" |
| 14:11.52 | nevcairiel | maybe you suck at understanding things |
| 14:12.04 | znf | nope, it's clear to everyone I talked to |
| 14:12.09 | znf | unless they fucked up the press release |
| 14:12.10 | nevcairiel | apparently not |
| 14:12.14 | nevcairiel | since you seem thoroughly confused |
| 14:12.14 | nevcairiel | :p |
| 14:15.49 | znf | it's unclear because they fucking announce this shit before they have the procedure tought out :-/ |
| 15:38.17 | *** join/#wowace Marcy_ (~textual@162-223-221-68.cpe.ngl.ninja) |
| 16:54.04 | *** join/#wowace EthanCentaurai (~EthanCent@host-78-147-99-38.as13285.net) |
| 17:12.24 | *** join/#wowace Seerah (~umsin@2601:3c2:101:320:64f9:56d6:4219:6d7f) |
| 17:15.43 | *** join/#wowace Sintacks (~Syntax@74-194-182-70.swbycmtk01.res.dyn.suddenlink.net) |
| 17:36.33 | Gnarfoz | and after the fact, as well :D |
| 17:40.02 | Gnarfoz | why are my soft-media keys so fucky right now |
| 17:40.27 | Gnarfoz | #a::SendInput {Media_Prev} |
| 17:41.20 | Gnarfoz | how does that result in https://hastebin.com/dahubuyana.css |
| 17:41.30 | znf | I need another car, Gnarfoz :( |
| 17:41.33 | Gnarfoz | I never pressed control |
| 17:45.24 | quiescens | o.O |
| 17:48.24 | Gnarfoz | apparently, AHK does that for me |
| 17:49.04 | quiescens | it presses control for you automatically? |
| 17:49.16 | Gnarfoz | yes |
| 17:49.16 | znf | Gnarfoz, have you figured out how to run AMR on Azure? |
| 17:49.26 | quiescens | maybe you secretly wanted to press control |
| 17:49.34 | Gnarfoz | running it on Azure can't be hard, the question is automation ;p |
| 17:49.42 | znf | that's what I mean |
| 17:50.08 | Gnarfoz | https://autohotkey.com/board/topic/20619-extraneous-control-key-presses-generated-by-or-hotkeys/ |
| 17:50.33 | Gnarfoz | it's intentional, to disguise the key prevent Windows from reacting |
| 17:51.02 | Gnarfoz | which doesn't quite explain why GPMDP doesn't react properly |
| 17:52.53 | quiescens | maybe we can just blame znf |
| 17:54.58 | znf | Gnarfoz, what was that thing called? |
| 17:55.05 | znf | you told me to use some batch thing on azure |
| 17:55.05 | Gnarfoz | which thing |
| 17:55.10 | Gnarfoz | it's called batch |
| 17:55.11 | znf | I don't remember the exact name |
| 17:55.15 | Gnarfoz | Azure Batch |
| 17:55.36 | Gnarfoz | it's the equivalent of AWS spot instances or GCP preemptible instances |
| 17:57.16 | Gnarfoz | (well, all of them work kind of differently, but it's the thing where instances are cheaper than usual) |
| 18:01.51 | znf | yeah, I don't understand that deployment crap at all :D |
| 18:05.16 | znf | also, by how I'm looking at it, you're supposed to have some monitoring/deployment script that runs the machines only when they have work to do |
| 18:07.56 | znf | azure is pretty much a clusterfuck that needs a training course just to figure out how to use it |
| 18:08.08 | Gnarfoz | same for AWS tbh |
| 18:08.17 | Gnarfoz | but I had one, so I knew roughly what to do |
| 18:09.02 | Gnarfoz | anyway, you're saying there's no way to manually spin up batch instances? that seems odd |
| 18:09.14 | znf | I can't found out how, yet |
| 18:09.31 | znf | find |
| 18:12.54 | znf | Gnarfoz, it's still showing me 112â¬/node/mo |
| 18:13.06 | znf | for 2 cores |
| 18:13.11 | Gnarfoz | embrace the $/h |
| 18:15.47 | znf | https://azure.microsoft.com/en-us/pricing/details/app-service/ |
| 18:18.56 | quiescens | why does everything you do sound so complicated znf ): |
| 18:19.03 | znf | it's not my fault |
| 18:19.04 | znf | it's Gnarfoz's |
| 18:26.28 | Gnarfoz | definitely not |
| 18:28.18 | znf | I've somehow managed to start a VM |
| 18:30.42 | znf | Task completed with exit code 'null' |
| 18:30.42 | znf | Essentials |
| 19:08.27 | Megalon | gz |
| 19:09.45 | *** join/#wowace dracula (~dracula@ip-217-103-124-30.ip.prioritytelecom.net) |
| 20:06.25 | *** join/#wowace dracula (~dracula@ip-217-103-124-30.ip.prioritytelecom.net) |
| 21:04.05 | znf | "dark tower" was nice |
| 21:29.58 | *** join/#wowace kroltan (~kroltan@200.194.235.247) |
| 21:37.47 | znf | Gnarfoz, I think I figured it out |
| 21:38.11 | Gnarfoz | now write a tutorial! :D |
| 21:38.35 | znf | you insane? >_< |
| 21:39.00 | znf | my VMs are in the state of "Starting" |
| 21:39.32 | znf | <PROTECTED> |
| 21:39.35 | znf | son of a bitch |
| 21:42.08 | znf | and they're still in starting |
| 21:42.13 | znf | jesus, this azure shit is so slow |
| 21:48.54 | znf | yeah, well, not quite |
| 21:49.00 | znf | I have managed to start instances |
| 21:49.13 | znf | but I can't seem to be able to actually start the application |
| 21:58.57 | znf | see, Gnarfoz |
| 21:58.57 | znf | $TargetDedicatedNodes=min(maxNumberofVMs, pendingTaskSamples); |
| 21:59.00 | znf | try to make sense of this |
| 21:59.23 | Gnarfoz | without context, that's rather pointless |
| 22:01.20 | znf | https://docs.microsoft.com/en-us/azure/batch/batch-automatic-scaling |
| 22:04.53 | znf | Gnarfoz, https://hastebin.com/recosuburi.makefile |
| 22:04.56 | znf | you think this makes sense? |
| 22:05.03 | znf | oh wait, |
| 22:07.30 | znf | ok, I think this makes sense https://hastebin.com/ezejopolav.makefile |
| 22:10.31 | znf | ie: if average CPU usage is over 10%, then start as many VMs as possible (which I think it's 4 right now for free accounts) |
| 22:10.36 | znf | else, kill the remaining nodes |
| 22:10.49 | znf | now if I could only figure the fuck out how to run an actual app on this thing |
| 22:11.28 | Gnarfoz | if it's a VM, how hard can that be? |
| 22:11.54 | Gnarfoz | but yeah, that script... thing looks terrible |
| 22:12.09 | Gnarfoz | can't you just give it a bash script to execute on startup |
| 22:13.32 | *** join/#wowace Megalon (~wig0r@193-81-178-119.adsl.highway.telekom.at) |
| 22:13.43 | Gnarfoz | (also, I think TargetDedicatedNodes should always be 0 unless you want to be poor real quick) |
| 22:14.13 | Gnarfoz | "Low-priority VMs are currently in preview, and are available only for workloads running in Batch." I think this might be half the problem right there. :D |
| 22:15.04 | znf | there's a limit of 8 cores for lowpriority |
| 22:15.10 | znf | that's why I'm running dedicated ones, too |
| 22:15.29 | znf | but I'm only running them when there are tasks to be run by AMR |
| 22:15.33 | znf | ie: no idling around for nothing |
| 22:16.38 | znf | the task command line is... hard to fucking figure out, really |
| 22:16.48 | znf | I tried doing it the most obvious way |
| 22:16.57 | znf | uploading an application (a zip of the amr client) |
| 22:17.03 | znf | and telling it to run amr.exe |
| 22:17.08 | znf | didn't work |
| 22:17.47 | znf | I tried writing a long linux command (or, rather, a simple script) that downloads the client and shit like that, didn't work |
| 22:17.48 | znf | :-| |
| 22:25.23 | Gnarfoz | https://hastebin.com/ofetequrof.bash is a trimmed down version of what I used on AWS |
| 22:25.30 | Gnarfoz | assumes it's Ubuntu 14.04 |
| 22:25.56 | Gnarfoz | (or rather whateveris in the 14.04 AMI on AWS) |
| 22:25.59 | znf | pft, you're actually creating the json :D |
| 22:26.23 | znf | I just created a tarball of a configured client :D |
| 22:26.26 | znf | and downloading that directly :D |
| 22:26.52 | znf | but yes, I tried that |
| 22:26.55 | znf | doesn't actually work :-| |
| 22:27.15 | Gnarfoz | the full version also adds a crontab entry that terminates the aws spot fleet if load15 goes below 1 |
| 22:28.01 | znf | but yeah, azure doesn't do this :-| |
| 22:32.57 | znf | http://i.imgur.com/BeGas5o.png |
| 22:33.06 | znf | I've no idea what this thing wants from me |
| 22:34.01 | znf | The following example shows a request body that could be used when automatic scaling is enabled. The formula increases the number of nodes that are used by 10% when the CPU usage in the pool for past 10 minutes stays over 70%. + |
| 22:34.07 | znf | totalNodes=($CPUPercent.GetSamplePercent(TimeInterval_Minute*0,TimeInterval_Minute*10)<0.7?5:(min($CPUPercent.GetSample(TimeInterval_Minute*0, TimeInterval_Minute*10))>0.8?($CurrentDedicated*1.1):$CurrentDedicated));$targetDedicatedNodes=min(100,totalNodes); |
| 22:43.12 | znf | Gnarfoz, ok, so my script actually works |
| 22:43.18 | Gnarfoz | good job \o/ |
| 22:43.20 | znf | not yet |
| 22:43.33 | znf | I've done curl https://my.script/sh | sh |
| 22:43.37 | znf | and it downloads it |
| 22:43.47 | znf | as I can see on my webserves logs |
| 22:43.50 | znf | but it doesn't actually run it |
| 22:43.51 | znf | ??!! |
| 22:44.12 | Gnarfoz | if it's like AWS, then it's probably running in a shell without a pty |
| 22:44.20 | Gnarfoz | hence my use of tmux |
| 22:45.24 | znf | I'm doing tmux too, obviously |
| 22:45.32 | znf | thing is - I can't find where the fuck it downloads it |
| 23:00.29 | Gnarfoz | curl "my.script/$PWD" :P |
| 23:00.46 | znf | Gnarfoz, what's a doubleVec |
| 23:02.15 | Gnarfoz | two... vectors? |
| 23:02.21 | Gnarfoz | how the hell would I know :p |
| 23:04.00 | znf | aaaaah |
| 23:04.08 | znf | you can avg() the result of the doublevec |
| 23:04.13 | znf | ok, that makes sense |
| 23:05.06 | znf | http://i.imgur.com/iTQoJgp.png |
| 23:05.52 | znf | there's a Sample every 30 seconds for CPU usage; so $CPUPercent.GetSample retries the samples of from TimeInterval_Minute * X minutes |
| 23:05.56 | znf | and returns a doublevec |
| 23:05.59 | znf | which you can avg() |
| 23:07.20 | znf | ok, scaling issues fixed |
| 23:07.30 | znf | *now* to figure out why this shit ain't starting |
| 23:12.21 | znf | 52.174.33.165 - - [08/Aug/2017:02:12:09 +0300] "GET /$PWD HTTP/1.1" 404 35639 "-" "curl/7.35.0" |
| 23:12.21 | znf | 52.174.33.165 - - [08/Aug/2017:02:12:09 +0300] "GET /run-amr.sh HTTP/1.1" 200 184 "-" "curl/7.35.0" |
| 23:12.24 | znf | son of a bitch |
| 23:39.57 | znf | Gnarfoz, I think I figured out auto-scaling, finally |
| 23:42.43 | znf | Gnarfoz, https://hastebin.com/jiqotidiwo.rb |
| 23:42.44 | znf | :-| |
| 23:43.27 | Gnarfoz | eh? :D |
| 23:43.33 | znf | my command lines was |
| 23:43.37 | znf | curl https:/// | sh |
| 23:43.56 | znf | apparently azure doesn't understand the concept of piping |
| 23:44.06 | Gnarfoz | maybe you want curl "https://somewhere" -O - | sh? |
| 23:44.36 | Gnarfoz | otherwise you're just piping curl's status output to sh (ok I guess if the pipe itself doesn't work... hm |
| 23:44.51 | Gnarfoz | maybe it's being run in powershell! ^^ |
| 23:46.09 | znf | http://i.imgur.com/YaMRM1V.png |
| 23:46.11 | znf | what the fuck |
| 23:46.48 | znf | wait, I have a better idea |
| 23:49.20 | znf | http://i.imgur.com/pw7klwT.png |
| 23:50.13 | Gnarfoz | you're on the wrong path, it seems |
| 23:50.58 | znf | gonna try adding quotes |
| 23:52.19 | znf | ok, I have ANOTHER idea |
| 23:52.33 | Gnarfoz | uhm |
| 23:52.43 | Gnarfoz | where are you putting your "command line" |
| 23:54.01 | znf | trying this now http://i.imgur.com/3MoNsZl.png |
| 23:54.29 | Gnarfoz | yes |
| 23:54.36 | Gnarfoz | https://docs.microsoft.com/en-us/azure/batch/batch-compute-node-environment-variables#Command-line%20expansion%20of%20environment%20variables |
| 23:54.50 | Gnarfoz | "The command lines executed by tasks on compute nodes do not run under a shell. Therefore, these command lines cannot natively take advantage of shell features such as environment variable expansion (this includes the PATH). To take advantage of such features, you must invoke the shell in the command line." |
| 23:55.01 | znf | sonofabitch |
| 23:55.08 | Gnarfoz | so whatever you do, you want to bash -c "blaaaah" |
| 23:55.29 | znf | yeah |
| 23:55.43 | znf | still only 11 clients connected |
| 23:55.48 | znf | but an improvement |
| 23:55.51 | znf | now I get the files created |
| 23:57.58 | znf | but tmux ain't running |
| 23:59.41 | znf | this really makes no sense, stupid microsoft :( |
