| --- Log opened Thu Jun 27 00:00:05 2019 |
| 00:04 | | Kindamoody is now known as Kindamoody[zZz] |
| 00:32 | <&ToxicFrog> | There we go. SHODAN now supports sharded SPUs. |
| 00:32 | <&ToxicFrog> | This should dramatically reduce hack overcommit. |
| 01:30 | | celmin|away is now known as celticminstrel |
| 01:42 | | abudhabi [abudhabi@Nightstar-7nkq9k.de] has quit [Operation timed out] |
| 01:52 | | abudhabi [abudhabi@Nightstar-7nkq9k.de] has joined #code |
| 01:52 | | mode/#code [+o abudhabi] by ChanServ |
| 01:59 | | abudhabi [abudhabi@Nightstar-7nkq9k.de] has quit [Ping timeout: 121 seconds] |
| 02:07 | | abudhabi [abudhabi@Nightstar-7nkq9k.de] has joined #code |
| 02:07 | | mode/#code [+o abudhabi] by ChanServ |
| 02:15 | | abudhabi [abudhabi@Nightstar-7nkq9k.de] has quit [Operation timed out] |
| 02:26 | | abudhabi [abudhabi@Nightstar-7nkq9k.de] has joined #code |
| 02:26 | | mode/#code [+o abudhabi] by ChanServ |
| 02:47 | | abudhabi [abudhabi@Nightstar-7nkq9k.de] has quit [Operation timed out] |
| 02:57 | | abudhabi [abudhabi@Nightstar-fd1.39v.247.94.IP] has joined #code |
| 02:58 | | abudhabi is now known as NSGuest38750 |
| 03:14 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has quit [Ping timeout: 121 seconds] |
| 04:53 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 04:53 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 06:59 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has quit [Connection closed] |
| 08:11 | | Kindamoody[zZz] is now known as Kindamoody |
| 10:07 | | jerith [jerith@Nightstar-ip7ar2.slipgate.net] has quit [Connection closed] |
| 10:13 | | VirusJTG [VirusJTG@Nightstar-42s.jso.104.208.IP] has joined #code |
| 10:13 | | mode/#code [+ao VirusJTG VirusJTG] by ChanServ |
| 13:09 | | celticminstrel [celticminst@Nightstar-6an2qt.dsl.bell.ca] has joined #code |
| 13:09 | | mode/#code [+o celticminstrel] by ChanServ |
| 13:18 | | celticminstrel is now known as celmin|away |
| 13:44 | | Emmy [Emmy@Nightstar-9p7hb1.direct-adsl.nl] has joined #code |
| 14:13 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 14:20 | <@TheWatcher> | # Can't do any forking work if there are no users to make forks for |
| 14:21 | <@TheWatcher> | >.> |
| 14:21 | <@TheWatcher> | Not swearing in work code, nope |
| 14:26 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has joined #code |
| 14:26 | | mode/#code [+qo Vorntastic Vorntastic] by ChanServ |
| 14:36 | | Kindamoody is now known as Kindamoody|out |
| 14:51 | <&ToxicFrog> | Implenting sharded SPU support has made the SPU and server upgrade code an unreliable nightmare. |
| 14:59 | <~Vorntastic> | Somehow I find that unsurprising |
| 15:15 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has joined #code |
| 15:15 | | mode/#code [+qo Vornicus Vornicus] by ChanServ |
| 16:09 | <&ToxicFrog> | I discovered this last night when I was too tired to fix it properly, so at the moment SHODAN just reboots completely after each server upgrade |
| 16:10 | <&[R]> | Could you make a work queue for each node, then have the update job just be added to the queue? |
| 16:11 | <&ToxicFrog> | [R]: the issue is that one node may be running arbitrarily many SPUs. Upgrading the node will restart all of them. |
| 16:11 | <&ToxicFrog> | But we notice that the node is ready for upgrade as soon as any of the SPUs on it reports in for a new work assignment. |
| 16:12 | <&[R]> | Hmm |
| 16:12 | <&ToxicFrog> | Immediately upgrading and rebooting the node when that happens means the other N-1 SPUs are still recorded as working on things, as far as SHODAN is concerned, which means subsequent work assignments will be wrong. |
| 16:13 | <&ToxicFrog> | (or rather: SHODAN records that the work is being done, but not which SPUs are doing it. So it knows which SPUs are being restarted but has no way of updating the task queue to take into account the fact that some tasks were interrupted and need to be retried, as opposed to just taking a really long time to complete) |
| 16:13 | <&ToxicFrog> | (Rebooting it fixes this because that drops the entire task queue and causes it to be incrementally rebuilt as SPUs report in) |
| 16:14 | <&[R]> | Do the nodes have any logic in them? Could you have them have an "waiting to upgrade" state where they'll eat the SPU "more work" messages? |
| 16:14 | <&[R]> | Then do the update once the SPUs are finished? |
| 16:14 | <&[R]> | Or even better, can you move SPUs to other nodes? |
| 16:19 | <@sshine> | - if ( $dest !~ /^(...|am|holgaard|morten)\@gnu/ ) { |
| 16:19 | <@sshine> | + # Agent forwarding is bad for security so we try to limit it |
| 16:19 | <@sshine> | + if ( ! ($server eq "gnu.jobsafari.dk" && $user ~~ $devs) ) { |
| 16:20 | <@sshine> | before, we limited SSH agent forwarding to developers who all had three-letter user accounts. now we have too many employees for that property to be true. |
| 16:21 | <@sshine> | except of course for those who worked here for 15+ years before the three-letter policy came into place. |
| 16:21 | <&[R]> | How is perl hooked into your sshd? |
| 16:21 | <&ToxicFrog> | [R]: the nodes have no logic in them, they're just hardware, and the SPUs themselves are very simple to maximize the number of them I can run. |
| 16:22 | <@sshine> | [R], it's a deployment script. |
| 16:23 | <&[R]> | Oh, so it generates a configuration to limit forwarding except for those users? |
| 16:23 | | * [R] is curious how you've got this setup |
| 16:24 | <&[R]> | Actually, I guess it doesn't matter, since I can disable agent forwarding in the key itself |
| 16:25 | <~Vornicus> | Having done the first thing in bitburner I cannot tell at all what the second thing should be |
| 16:25 | <@sshine> | I guess that's even safer. |
| 16:27 | <@sshine> | [R], looking into it, our practice is not recommendable. it's basically a perl script that propagates a set of SSH public keys that are hardcoded in the file to a bunch of servers. if the script screws up too much, we risk wrecking SSH access. |
| 16:28 | <&[R]> | Yeah, I have a host that generates the keys and signs them (this also lets me make sure that passphrases are used and meet minimum requirements), the signing process lets me also specify rules on the key. For example, I can make a key that can only be used to run a specific command. |
| 16:28 | <&[R]> | Then on the hosts, they only allow keys that are signed by the cert-key |
| 16:29 | <&[R]> | This also makes enabling access to a system fairly simple. Just make the user, the key will automatically let that user in. |
| 16:30 | <&ToxicFrog> | Vornicus: my progression after the tutorial was something like: |
| 16:30 | <@sshine> | there's a lot of deployment, containerization, CI etc. I'd like to dig deeper into, but my current job is built from brambles. |
| 16:31 | <&ToxicFrog> | - program to automatically weaken/hack/grow the server it's running on, while keeping money increasing up to the server's max |
| 16:31 | <&ToxicFrog> | - program to automatically scan the network and deploy the first program onto any machine I have root on |
| 16:31 | <&ToxicFrog> | - program to automatically scan the network and get root on things |
| 16:31 | <&ToxicFrog> | - hacknet node auto-purchaser/upgrader |
| 16:31 | <&ToxicFrog> | - sudden realization that hack() doesn't need to be called on localhost, write an RPC library and then SHODAN |
| 16:32 | <&ToxicFrog> | - fantastic wealth, start slowly grinding towards the requirements in fl1ght.exe across several reset |
| 16:33 | <&[R]> | Anyways, BBL |
| 16:33 | <&ToxicFrog> | Oh yeah, and once you have distributed hacking up and running, purchasing/upgrading dedicated servers is a huge speed boost once you can afford it -- definitely write a program for that, saves you having to fly around to different computer stores in person. |
| 16:46 | <&ToxicFrog> | Vornicus: more generally, as your hack and money increase you'll get more messages from jump3r and hacking groups and invites from various factions. The factions give you tasty, tasty augs. The messages will point you at more servers; hacking them *manually* will generally get you a new faction invite, although some just contain interesting documents or coding contracts or the like. |
| 16:52 | | jerith [jerith@Nightstar-ip7ar2.slipgate.net] has joined #code |
| 16:53 | | mode/#code [+ao jerith jerith] by ChanServ |
| 17:38 | | Pinkhair [user1@Nightstar-g7hdo5.dyn.optonline.net] has quit [Ping timeout: 121 seconds] |
| 18:13 | | Vorntastic [uid293981@Nightstar-6br85t.irccloud.com] has quit [[NS] Quit: Connection closed for inactivity] |
| 18:16 | | gnolam_ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has joined #code |
| 18:17 | | gnolam__ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has joined #code |
| 18:19 | | gnolam___ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has joined #code |
| 18:20 | | gnolam [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has quit [NickServ (RECOVER command used by gnolam___)] |
| 18:20 | | gnolam___ is now known as gnolam |
| 18:20 | | mode/#code [+o gnolam] by ChanServ |
| 18:20 | | gnolam___ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has joined #code |
| 18:21 | | gnolam_ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has quit [Ping timeout: 121 seconds] |
| 18:22 | | gnolam_ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has joined #code |
| 18:22 | | gnolam__ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has quit [Ping timeout: 121 seconds] |
| 18:24 | | gnolam [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has quit [Ping timeout: 121 seconds] |
| 18:24 | | gnolam_ is now known as gnolam |
| 18:24 | | mode/#code [+o gnolam] by ChanServ |
| 18:25 | | gnolam___ [lenin@Nightstar-hfrbpd.cust.bahnhof.se] has quit [Ping timeout: 121 seconds] |
| 19:56 | | Kindamoody|out is now known as Kindamoody |
| 20:42 | < Mahal> | https://twitter.com/jaredpalmer/status/1142470313592143872 |
| 20:44 | <@Tamber> | ha |
| 21:03 | | NSGuest38750 is now known as abudhabi |
| 21:03 | < abudhabi> | What's 10x? |
| 21:21 | | catalyst [Jessikat@Nightstar-5dv16h.cable.virginm.net] has joined #code |
| 21:34 | <@ErikMesoy> | Engineer that is ten times as skilled/productive/mythical as an ordinary engineer. |
| 21:42 | <&McMartin> | People disbelieve in the existence of 10x engineers but have no trouble believing in 0.1x ones~ |
| 21:42 | <~Vornicus> | 10x engineers are what pessimists believe in |
| 21:42 | <~Vornicus> | 0.1x engineers are what optimists believe in |
| 21:44 | <&McMartin> | The intro to The Mythical Man-Month had this nailed, I think |
| 21:44 | <~Vornicus> | Having not actually read it |
| 21:45 | <~Vornicus> | You will have to explain |
| 21:45 | <&McMartin> | Yeah, it was not a main thesis, but it is long so will take more than one line |
| 21:45 | <&McMartin> | You have programs |
| 21:45 | <&McMartin> | You have programming systems, which is a program that can be maintained and extended later by strangers |
| 21:45 | <&McMartin> | And you have programming products, which are properly tested, scalable, run on systems that aren't your laptop or hardware equivalent to same, etc. etc. etc. |
| 21:46 | <&McMartin> | And you have programming systems products, which are both of the above and "the truly valuable thing" |
| 21:46 | <&McMartin> | Draw that on a 2D grid with programs in the upper left and programming systems products in the lower right |
| 21:46 | <&McMartin> | Taking one step down or right increases the complexity/cost of the task by 3x |
| 21:46 | <&McMartin> | 3 squared is... about a 10x difference in expected productivity |
| 21:48 | <&McMartin> | The question being answered there was, at the time, asked with something like "Okay, so, we've all heard the stories of major breakthroughs in software developed by, like, two dudes in a garage. Companies aren't idiots, so: why isn't two-dudes-in-a-garage being replicated on an industrial scale?" |
| 21:52 | <~Vornicus> | Having done mostly the top left and a little in the bottom left, yeah, okay |
| 22:26 | | Vornicus [Vorn@ServerAdministrator.Nightstar.Net] has quit [Connection closed] |
| 22:29 | | catalyst [Jessikat@Nightstar-5dv16h.cable.virginm.net] has quit [Connection reset by peer] |
| 22:56 | <&McMartin> | The top comment on this article is completely unwarranted and unfair and I can't stop laughing at it anyway https://arstechnica.com/gadgets/2019/06/gmails-api-lockdown-will-kill-some-third-party-app-access-starting-july-15/ |
| 23:07 | <@Tamber> | ha |
| 23:29 | | himi [sjjf@Nightstar-v37cpe.internode.on.net] has quit [Connection closed] |
| 23:39 | | Lamb3 [Dawg@Nightstar-52ic2g.wechall.net] has quit [Connection closed] |
| --- Log closed Fri Jun 28 00:00:06 2019 |