[02:15] Hopefully a silly question [02:15] How do I make my DNS server accessible outside of the host that it's on? [02:15] having issues hitting it from a different machine, and as far as I'm aware Ubuntu 18.04 doesn't have a firewall by default... or so I thought... [02:16] I've checked ufw is disabled [02:17] did you check what IP addresses your dns server is supposed to be listening on? [02:17] if you're running it on a cloud provider, did you let through both tcp and udp 53? [02:17] ...actually. I'm being stupid. [02:17] I'm trying to do this on an Ubuntu Desktop machine, and that is almost certainly causing a conflict. [02:17] I might try rebuilding on Ubuntu Server instead [02:18] why? [02:18] why was I building on Desktop? [02:19] no, why would it make a difference? [02:20] it's just more complicated I guess? [02:20] anyways. The box is running on my machine, in VirtualBox [02:21] has the IP 10.21.30.2, and other machines can ping it just fine. I can hit port 53 on the machine via telnet and get a response, but not from another machine [02:26] how about other services on the machine? sshd? web? [02:31] hmm, desktop and server are the exact same thing [02:31] define another machine [02:31] in virtualbox or actually a physical machine [02:32] and always try to test with ping also :) [02:36] well, server won't install with networkmanager as the default netplan renderer, but once you've got an ip address, they'll be pretty similar :) [02:37] I dunno why netplan was put into place, so many things it doesn't support, again [04:04] patdk-lap: I'm just trying it as a hunch... worst case I learn something [04:05] And I meant another VM on the same network, which can ping, SSH and otherwise contact the DNS server, just no DNS... [04:06] I should say DNS *host* === mIk3_09 is now known as mIk3_08 [08:18] sahid: I'm going to start on neutron* and networking* [08:37] jamespage: ack [08:41] sahid: making a fix to openstack-pkg-tools to restore the understanding of git snapshots when generated OSLO_VERSION [08:41] that was lost in the last sync from Debian [08:51] jamespage: how i determine the next version? [08:52] sahid: I do previous release major version +1 [08:52] so for neutron [08:52] example for aodh, stein version is 8.* so i imagine for Train it's 9. [08:53] 14.0.x is stein; train is 15.0.0~b1 for snapshots (+ git references and date) [08:53] sahid: yep you got it [09:37] jamespage: how do you handle (build-)depends? [09:47] jamespage: if you have a moment at some point to validate https://code.launchpad.net/~sahid-ferdjaoui/ubuntu/+source/aodh/+git/aodh [09:47] at least to ensure that i'm doing it in the right way [09:56] sahid: https://git.launchpad.net/~sahid-ferdjaoui/ubuntu/+source/aodh/commit/?id=a58ef18fd58dc74a67fc8cefc9d28e87fccdcaa5 [09:57] when you version depends like this please add to both Build and Runtime Depends - the python3-aodh package should mirror the source package build depends versions [10:04] sahid: other than that LGTM - are you build testing? I generally use a PPA and throw stuff at it until it works [10:04] https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3690/+packages [10:04] example [10:24] sahid: https://launchpad.net/ubuntu/+source/openstack-pkg-tools/99ubuntu1 should re-enable the correct setting of OSLO_VERSION when using our git snapshot versioning semantics === kuato is now known as dryliketoast [10:38] Can someone comment on the accuracy of this statement, it is from IBM: https://pastebin.com/VqHzmTYU [10:45] Ussat: this may refer to https://wiki.ubuntu.com/BionicBeaver/ReleaseNotes#A32-bit_PowerPC_Support_Dropped - see https://ubuntu.com/download/server/power for supported POWER platforms. [10:46] OK...thats to bad [10:48] 32-bit POWER has kind of reached the end of its lifespan in general, though. [10:48] Sure......I did not post the entire email...... [10:48] this is referencing 64bit ppcle [10:49] specifically, bnot refering to the ppcle platform, but the PowerVM Hypervisor [10:49] PowerVM enables DLPAR and Logical Partition Mobility [10:51] Here is more detail: https://paste.centos.org/view/bea75610 [10:52] god dammit stripped the http [10:52] sigh [10:53] There: https://pastebin.com/Tt3QZEET [10:53] Thats the entire email minus identifying info, like names :) [10:54] Note, PowerVM is different than PowerPCle [10:54] PowerVM is a hypervisor that runs on PowerPCle [10:54] and I have the 18.04ppcle iso [11:10] PowerVM does not seem to be specific to 32-bit POWER platforms, support for which got removed between 16.04 LTS and 18.04 LTS. so i'm not sure what they're referring to. [11:12] for all i can tell there was never direct "support for PowerVM" by Ubuntu [11:12] OK, I will try to clarify with IBM, thanks [11:13] so i'm not sure what they're referring to by stating "Ubuntu effectively stopped support for Ununtu PowerVM's with Version 16" [11:14] Ussat: Note that i'm just a volunteer, not a Cannical spokesperson nor Ubuntu developer. [11:14] I will try to get clarification from IBM on this, thanks [11:14] NP, all good and appreciate the info [11:15] I may call Canonical directly later, thanks again [11:36] They may be referring to the fact that there's no certification beyond 16.04 for PowerVM [11:47] Thats probably it, I have an email to my IBM rep [11:47] TJ-, if thats the case, its unfortunate [11:47] I also have an email to Canonical [11:48] It really limits my choice to RHEL or Cent on PowerVM [12:44] patdk-lap: such as? (things that netplan does not support) [12:48] dummy interfaces [12:49] atleast that I ran across the first installed I did that had netplan [14:16] sahid: do we need a futurist version bump? [14:16] just looking at your build errors [14:16] jamespage: the OSLO fix you did resolved an issue with oslo_upgradechck? [14:17] jamespage: i need to check that, currently i just reported the issue [14:19] jamespage: python3-futurist is in eoan-proposed [14:21] i mean the 1.8.1 version needed [14:21] perhaps the buildroot that i use in my ppa should be based on proposed? [14:51] sahid: yes - you can tweak that in the ppa configuration screen === waveform_ is now known as waveform [15:44] Hi @TJ- [16:01] jamespage: the ppa is rejecting my new uploads, i guess is because the version does not change, any idea how i can force? [16:06] sahid: you can't [16:06] once a version is uploaded its burned [16:06] sahid: that's why I used a generated version ID - https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/3690 [16:07] sahid: its just a wrapper around backportpackage - https://paste.ubuntu.com/p/KrJrfpFTMx/ [16:07] I build the source package with my intended upload to ubuntu version, and then backportppa -u -d eoan -y .dsc [16:07] the version has timestamp in it so always moves forward correctly [16:08] sahid: PPA's work just like the main archive does from this perspective [16:09] ack, wil try to understand all of that, thanks a lot [16:10] sahid: you can delete the current set of packages from the PPA, and then use that script to upload new ones [16:10] that should work [16:13] I've got an apt-get upgrade process running that I don't recall starting myself. Is there some way I can check its current status, make sure it's actually doing something? [16:14] I imagine this is just some automagic security update request or something so as long as it's not a zombie process I'm okay with letting it finish first. [16:16] ps aux says "jun05" yeah it's probably a zombie nvm lol [17:08] three cheers for openssl 1.1.1! hip hip, hooray! hip hip, hooray! hip hip, hooray! [17:09] Ooh, TLS 1.3, 0-RTT, SHA3, nice! === MassDebates_ is now known as MassDebates [18:39] Hrm. So, the topic links https://help.ubuntu.com/16.04/serverguide/, but https://help.ubuntu.com/18.04/serverguide/ is available. [18:39] Just saying. [18:48] it should probably point to https://help.ubuntu.com/stable/serverguide/ instead [19:03] Even better. [19:29] So, I'm curious... Is there a trick to getting a VM to support S3/S4? [19:30] libvirt/virsh, I see: "error: internal error: S3 state is disabled for this domain" [19:34] mason, what's S3 in this context? [19:34] lordcirth: Sleep state. As in, "virsh dompmsuspend --target mem foo" [20:14] anyone know anything about STunnel? [20:15] teward: yes :) [20:15] teward: I used to use about 20 years ago with Windows :p as well as on Linux [20:17] :P [20:17] trying to use it to get a secure tunnel to Redis but... [20:17] getting this on client side: SSL_connect: 14212044: error:14212044:SSL routines:tls_construct_ctos_early_data:internal error and this server: SSL_accept: 140940F4: error:140940F4:SSL routines:ssl3_read_bytes:unexpected message [20:17] makes me think TLS 1.3 is at fault [20:17] but this works between containers so IDK [20:18] different ssl libraries at either end? [20:18] early_data sounds like 0-RTT [20:20] TJ-: 1.1.1 on both sides [20:20] unless stunnel4 needs a rebuild after the OpenSSL backport on 18.04 [20:21] but the odd thing is this works FINE on stunnel4 between containers [20:21] and the WEIRD part is it's accepting the connection from remote THEN barfing [20:21] * teward grumbles. [20:22] is "ssl3_read_bytes" still used with TLS 1.3? [20:22] no idea, but this is the OpenSSL error it triggers [20:23] teward: have you done a test with openssl s_client ? [20:23]

fwiw [20:23] mason: hah, thanks [20:25] teward: the failing connection is stunnel4<>stunnel4<>redis or stunnel4<>redis ? [20:25] Ah, didn't work. Still "S3 state is disabled" [20:25] sigh [20:25] TJ-: python Redis client (PLAIN) <> stunnel4 CLIENT <> stunnel4 SERVER <> redis SERVER [20:25] mason: not sure that interests you but there is 'virsh suspend $VM' that pauses the VM [20:25] breaks between the two stunnels when actually going between networks [20:25] BUT [20:26] I'll figure it out anyway. Tracking a suspend/resume bug, and it'd be a happier picture if I can debug it on a VM [20:26] same setup worked FINE in same network between containers running the same OS (18.04) [20:26] teward: so in all cases we're dealing with 2 stunnel4 instances [20:26] TJ-: correct [20:26] sdeziel: It might help. Unsure. This manifests as an ACPI bug on real hardware. [20:26] but it doesn't work over the Internet but DOES through the local containers subnet without going to the Internet or passing between network layers like that [20:26] *shrugs* [20:27] mason: if you want to debug a suspend/resume bug, I doubt cause 'virsh suspend' seems to just send a stop command to the QEMU [20:27] oh, it overwrote the change on shutdown :P me-- [20:27] :( [20:27] oh hm hang on [20:27] i think i might've broke something here [20:27] teward: it can happen if the ports are being scanned [20:28] TJ-: this can also happen if it doesn't get a cert [20:28] \o/ [20:28] ssl handshake failure sclient [20:28] 1 moment [20:28] teward: I thought you said it had connected? [20:28] TJ-: it had guess I did a stupid somewhere [20:28] hang on [20:29] mason: you changed a manual config which got replaced? I had that happen to me earlier [20:30] TJ-: My mistake was changing the config before killing the VM. [20:30] mason: yeah, annoying when it writes the existing in-memory config out after you've edited it :D [20:30] changing it afterwards worked [20:30] yeah [20:30] spend time wondering how it got unset :p [20:30] YEP [20:33] We can get a vagrant VM to sleep. Odd. [20:34] sing it a sweet little song, rock it back and forth.. [20:36] heh [21:07] TJ-: i think i failed in cert config [21:07] but i will test at home :p [21:59] hello [22:02] Can someone help me with configureing ubuntu server with my network card? i am only familler with linux gui, im stuck at the installer becase its only looking on eathernet [22:08] Nikita790: what release of ubuntu? what are you trying to do? where are you stuck? [22:10] sorry [22:10] i am used to discords bleep msg sounds [22:11] Basicly im trying to install 19.04 and i am stuck at the network connections setup, it will not see my network card, it only sees my eathernet [22:11] How can i get it to see my network card? [22:11] so, you want wireless networking? [22:11] yes [22:12] find out what adapter, lspci would tell [22:12] and ifconfig would show more.. [22:13] one second [22:13] btw one must have a reason not to use 18.04 LTS .. [22:14] oh [22:14] i just chose 19.04 becase i tought it would be fastest [22:14] should i burn a 18.04 and use thatt instead? [22:15] yes, preferrably [22:15] ok i will do that right now [22:16] stable and well tested. and lots of guides only handle 18.04 [22:17] the card is a linksys 2.4g wmp54G [22:17] i know it works on lubuntu 19.04 [22:18] i am burning dvd now [22:19] thank you so much [22:20] no usb memory sticks? [22:21] no [22:21] i do not have one [22:22] sorry, i am not used to checking my irc becase im used to discords msg sounds, sorry [22:22] Nikita790: don't worry about it, people come and go all the time on irc [22:23] ok the iso is almost done downloading [22:23] then i can burn, hopefully it will recognize my linksys 2.4g wmp54G PCI wifi card [22:24] ok its burning at 8x speed [22:26] oke, after installing, use wired networking >> To use wifi with netplan.io and systemd-networkd, you need to manually install the wpasupplicant package. It is not automatically installed as a dependency since wifi support is optional on servers. [22:27] ok... thank you [22:27] I will do my best [22:30] ok [22:30] so [22:30] does usb eathernet work for the installer? thats all i have acces too [22:32] depends [22:32] i am going to try to use the ios hotspot one [22:32] thats all i have acces to other then wifi [22:32] if your usb network adapter has a linux driver in the kernel or not, likely does [22:32] i will try [22:33] it just gets so iffy, as it can take a few years sometimes for new device chipsets to get drivers into the kernel [22:33] ok i hope my iphone has it, im gonna use the hostspot [22:33] im booting up server 18.04 now [22:34] TJ-: well it 'connects' but... this is now what s_client shows: SSL_accept: 14201076: error:14201076:SSL routines:tls_choose_sigalg:no suitable signature algorithm [22:34] :| [22:35] looks like the key exchange algo lists are different? [22:35] possibly [22:35] teward: or is it the certificate signing algo its on about? [22:35] but that's s_client -> the stunnel4 server where redis is [22:35] directly [22:35] can you tell from the debug point [22:35] TJ-: can't tell [22:36] s_client can have very very verbose debug logging, which might help figure out which stage it is at [22:36] well if i drop to TLS1.2 [22:36] it says no shared cipher [22:36] which is a different error :| [22:36] this is annoying me [22:37] i wish there was a version of ubuntu server with all the ubuntu desktop wifi drivers baked in [22:37] teward: bingo "DSA certificates are no longer allowed in TLSv1.3. " [22:37] i'm not using DSA [22:37] they're RSA [22:37] and i'm forcing 1.2 now [22:38] bleh might blast this config and start over [22:38] see if i did something wrong in the config [22:38] teward: hmmm, see https://www.openssl.org/blog/blog/2017/05/04/tlsv1.3/ where they show that exact error [22:38] Nikita790: if you've got monitor on the computer you could probably install the ubuntu desktop and just remove the packages you don't want [22:38] TJ-: going to blast the configs and start over [22:38] with something that 'just works' to start [22:38] then try and add auth, etc. [22:38] * teward goes to copy directly the configs from the containers [22:39] Sarnold i used to do that but i heard it was very bad for the preformace [22:39] and i need max preformace becase this is a low end pc thats gonna host a game server [22:40] Nikita790: there's not much difference between the two, beside the desktop version installs a GUI and uses networkmanager to configure networking [22:40] you could easily uninstall both those if you wished [22:40] wow, ok i might keep my lubuntu installation then [22:40] i just cant get the bloody screensaver to be disabled lol. thank you so much for your wisdom [22:41] thanks [22:44] heh if that's the problem with your existing system I'm sure there's a solution of some sort :) dpkg -l '*screen*' might be a good start [22:46] oh yes i used screen :D i just heard that a entire gui killed the servers preformace [22:48] it depends what it's doing, how much GPU vs CPU vs memory it takes up, etc.. [22:48] the dpkg -l '*screen*' is to try to figure out what screensaver lubuntu might be using. it might be as easy as apt-get purge :) [22:52] TJ-: huh, you know what.. [22:52] this might mean the ssl-cert package that generates snakeoil certs needs updated [22:52] sarnold: ^ [22:52] teward: DSA cert? [22:52] TJ-: not 100% sure but I'd like to FORCE it to use RSA [22:52] checking now [22:53] there's an openssl.cnf in stunnel's /usr/share/doc/stunnel4/examples/ too, which might need looking at (it's for generating certs) [22:53] huh nope it's an RSA cert [22:53] TJ-: i was just trying to PSK the thing [22:53] looks like the system is weird :| [22:53] works FINE now i think [22:53] but i'll have to add PSK stuff in again for testing [22:54] OH i was told if i use the alternative iso i can use wifi [22:54] cert auth is even more painful [22:54] im trying that [22:54] so PSK for now with LONG keystrings [22:55] I never had a problem with certs; I use a USB Armory for issuing/signing certs [22:55] TJ-: yep working now. [22:55] TJ-: I use XCA but it's unclear what certs go where, and /usr/share/doc/stunnel4/* doesn't seem to exist [22:55] or at least it DIDN'T [22:55] :| [22:55] now it's here [22:56] i can generate the cert inside XCA now though [22:56] now that i now what extensions it needs xD [22:56] :) [22:56] but what I really need is client cert auth working [22:56] and I can't find examples of that [22:58] teward: https://launchpad.net/ubuntu/+source/ssl-cert/+changelog .. most recent change from 2017 .. since 1.1.1 is in cosmic, disco, eoan, without trouble, I'm guessing it's probably not in immediate need of update.. [22:58] sarnold: yeah i poked it's an RSA cert [22:59] i think SOMETHING was just fubar with the cert when being parsed, redid the cert by hand and made a selfsigned and it "just worked" [23:00] *shrugs* [23:01] :/ [23:02] :D [23:10] Selfsigned: Username checks out. [23:16] TJ-: OK so... [23:16] can't use SSL with PSK [23:16] so i'll have to just use a redis auth PW then [23:16] which i should do anyways lol [23:17] sarnold: is there any guide for converting a sysvinit into a SystemD unit? [23:19] teward: this is a nice overview https://wiki.ubuntu.com/SystemdForUpstartUsers [23:19] teward: (not of the exact sysv->systemd, but systemd in general) [23:20] yeah i'm going to futz with the stunnel4 package locally to see if I can't SystemD unit the entire thing [23:20] get it off the older methods [23:20] going to be an evil project but xD [23:20] tired of using sysvinit evil [23:23] yeah; I don't love systemd, but sysv-init isn't my idea of great either :) [23:25] teward: there's too much documenation for systemd unit files, and throwing you into the docs feels mean.. but if you've got a cause to run one stunnel4, you probably have cause to run several of them. and they might be similar enough to justify learning the 'template' support, mentioned on https://www.freedesktop.org/software/systemd/man/systemd.unit.html [23:25] (and the manpages, of course, but the hyperlinked ones are actually kind of nice) [23:26] sarnold: true. but I'm curious why we don't unzip and copy the sample config into /etc/stunnel/ [23:26] because it's weird, there's LITERALLY no config examples there by default [23:26] it's all dug deep in /usr/share/doc/... [23:26] maybe that's normal but eh [23:26] teward: yeah I always prefer having example configs in /etc .. but I can kind of understand the folks who want /etc to say what's *different* about the machine in question. it's weird. [23:28] i looked at the stunnel config example though [23:28] EVERY unit in its example is commented out [23:28] and /etc/default/stunnel4 has to have ENABLED=1 to actually start [23:28] so IDK [23:28] it'll be a long term project to SystemD-ify stunnel4 [23:29] or ignore the configs / initscript shipped in the packages? [23:30] lol indeed [23:30] sarnold: well i would want to ship it as by default NOT enabled [23:30] is that even doable in the package policy to autodisable the service at install? [23:30] because it would NEED configured to even run [23:30] per my testing at least [23:30] no config, megaerror [23:31] teward: good question. it's my understanding that systemd comes from the land of 'installed packages don't automatically run anything'.. I'm not sure how well that'd fit in debian [23:31] well i an't touching Debian with a fifty foot pole [23:31] :p [23:31] which reminds me [23:31] I still need to distropatch NGINX Eoan [23:31] with that PIDfile handling thing === BrianBlaze420 is now known as brianblaze [23:48] I'm trying to set up an iSCSI target with targetcli, I'm confused why various tutorials, and even the manpage refer to a systemd unit I don't have and cant find any info about how to install. Eg: "$ sudo systemctl enable target.service" [23:48] Does anyone know what package provides that? [23:48] http://manpages.ubuntu.com/manpages/bionic/man8/targetcli.8.html [23:59] teward: I can systemd-ify stunnel if you want the package updating; it doesn't require much at all