[00:00] what do you have in /var/log/mysql* ? [00:00] when you last asked the same question in #ubuntu it was suggested you take a look at https://dev.mysql.com/doc/refman/5.5/en/problems-connecting.html [00:00] I tried to find this file /var/run/mysqld/mysqld.sock and it doesn´t exist [00:00] that's because the server isn't running [00:00] I did create it and change permission but when I restart mysql it cleans up the file [00:01] it's not just a missing file, that's a unix socket [00:01] it's one way mysql clients can talk to the server [00:01] your problem is before that [00:02] tomreyn: I had my irc client set to not receive channel massages I guess [00:03] I was receiving some spam and I needed to set my client and I think I changes something wrong [00:04] ahasenack: I have error.log but nothing inportant on it and it won´t change in restart [00:05] ahasenack: if you think it can help I can paste it somewhere [00:05] what about "sudo systemctl status mysql"? [00:06] supercool: ok [00:06] you should probably post my.cnf and package information on the mysql server used, too [00:07] ahasenack: service mysql status == "* MySQL is stopped." [00:07] supercool: there is usually more [00:07] supercool: how about this [00:07] sudo systemctl start mysql [00:07] then [00:07] sudo systemctl status mysql [00:07] and paste /var/log/mysql/error.log [00:08] ahasenack: I don´t have systemctl here. Can I use service instead? [00:09] supercool: what's your ubuntu release? [00:09] 18.04 I think [00:09] then you have systemctl [00:09] How do I find it? [00:09] "lsb_release -ds" reports your ubuntu version [00:09] which systemctl [00:10] No, I mean the ubuntu release [00:11] see one line above what you read last [00:11] I don´t have lsb_release [00:12] Maybe because I install it on a docker guest [00:12] cat /etc/issue [00:12] and that might be related to your mysql problem [00:13] * ahasenack -> bed [00:13] good luck [00:13] ahasenack: thank you! [00:14] Ubuntu 18.04 LTS \n \l [00:14] tomreyn: thank you! [00:14] so it's not fully up to date. [00:15] install all pending updates, see if this helps [00:15] bash: systemctl: command not found [00:15] well, your system is not properly installed [00:15] how did you install it? [00:16] I used docker [00:17] the command was apt-get update [00:17] I think I need a apt-get upgrade there [00:18] Do you want to see the Dockerfile used to generate the image? [00:18] "i used docker" as a response to the question "how did you install [Ubuntu]" is like saying "i use a computer" in response to the question of how you installed your OS. [00:18] tomreyn: sorry, I don´t understand your question then. [00:19] What do you want to know with how did you install it? [00:19] well docker is not only but mostly an einvironment where something can operate in. it doesn't explain how the stuff that operates in it is setup. [00:20] okay, show the dockerfile [00:20] Docker has a Ubuntu repository with a installer image. I did build a image of it on my computer [00:21] were there any errors when you built it? [00:21] Man, sorry. Docker has a repository with the official Ubuntu-server image on it. [00:21] You just use it to remote install it on your computer. [00:21] what you have there is not a proper 18.04 installation, i'm surprised it boots at all. [00:22] I think it is a very basic install made just to boot [00:22] Then you add apt as you wish [00:22] This is the idea [00:23] can you run "dpkg -l 'systemd*'" on this docker iinstance? [00:23] yes, it is possible [00:24] Could you point me a paste bin where I can put it to you please? [00:24] okay, it should generate some output, tell me the lines which start with 'ii' [00:24] !paste [00:24] For posting multi-line texts into the channel, please use https://paste.ubuntu.com | To post !screenshots use https://imgur.com/ !pastebinit to paste directly from command line | Make sure you give us the URL for your paste - see also the channel topic. [00:25] Here we go: https://paste.ubuntu.com/p/4wfNJSxqzq/ [00:25] I don´t think that was useful at all [00:26] okay, so that's ubuntu 18.04 without systemd. not a proper installation. i can't help you with this. [00:26] I can install systemd if you wish [00:27] that's not the point. neither ubuntu desktop nor ubuntu server would look like this after installation. i do not know what this system is, but it is not something i am into, and so i cannot help you with it. [00:27] https://paste.ubuntu.com/p/FXkYSsFD4Q/ [00:28] yes, that's what it would normally look like [00:28] o/ [00:28] i can help you install mysql on an ubuntu server installation, but that's not what you have there. [00:28] maybe it's ubuntu core or something, i have no experience with this. [00:29] Well if there is any tool not present I can go installing it [00:29] Untill we figure whats goin on [00:29] until* [00:29] i'm not going this route, this is just try and error, not knowing what you work with. [00:30] a waste of time on your and my part [00:30] Don´t fell afraid [00:31] We can do it [00:31] the only thing i'm afriad of is wasting more time on this. good luck! [00:32] Alright. Thank you anyway tomreyn! [00:33] welcome. please be sure to point out that you're experimenting with docker early in the conversation next time. [00:36] tomreyn: Ok. === JanC_ is now known as JanC [04:49] good morning [04:54] ahasenack: 3:38 my time the importer completed the catch up === beatzz is now known as elsheepo [05:12] MSSQL is being really fickle tonight and I'm just considering just throwing it into a VM and never think about it again. [05:13] Really hates ZFS with a burning passion. [05:13] So might as well bubblewrap it inside a VM and avoid that trouble. === techmagus_ is now known as techmagus [05:57] Good morning [07:00] I was about to send an message mentioning that another server is going to have to handle requests for something and they might need to change settings. And then realized 'Oh I could just forward the port to the other server and no one has to do anything' [07:00] Yay for...laziness. [07:00] Or something. === SynchronE is now known as aleks_bogdanov === m1dnight_ is now known as m1dnight [08:35] debian-security-support exists in ubuntu, but I'm pretty sure it's useless. Is that right? Is there an Ubuntu equivalent? (I think it used to be called update-manager-core back in 2012) [08:36] update-manager-core is already installed on this host, so I'll call that "good enough" for today. [09:05] Hrm, why does unattended-upgrades in Ubuntu 16.04 still default to only Debian origins? [09:05] Origins-Pattern "origin=Debian,codename=${distro_codename},label=Debian-Security"; [09:06] That's what dpkg-configure says, but "apt-config dump" disagrees.... [09:24] so how do you configure a dummy0 interface on ubuntu server 18.04 using netplan [09:25] not only configure, but create as well [09:25] TvL2386: I'm not familiar with netplan. Have you already tried the obvious - check /usr/share/doc/netplan, man netplan ? [09:25] yep [09:26] It seems there's no such feature at the moment [09:26] https://bugs.launchpad.net/netplan/+bug/1774203 [09:26] Launchpad bug 1774203 in netplan "support for dummy interfaces" [Undecided,New] [09:26] the netplan I see in Debian seems to be for people doing finger(1) at MIT in 1987 [09:26] meh.... [09:27] what do you mean twb? [09:28] Ahhh, http://bugs.debian.org/882661 [09:28] Debian bug 882661 in wnpp "ITP: nplan -- YAML-based network configuration tool" [Wishlist,Open] [09:28] I'm not really a fan of netplan [09:28] cf. https://manpages.debian.org/netplan [09:29] oh lol :) [09:29] Grr, manpages.ubuntu.com doesn't do TLS, and requires javascript or something [09:30] Here we go: http://manpages.ubuntu.com/manpages/bionic/en/man8/netplan.8.html [09:30] so far my experience with netplan: pre/post hooks are not implemented and the workarounds are not working, bonding interfaces aren't removed when you remove them from your yaml file, dummy is not available [09:30] Oops that's still the silly one [09:31] is there a way to revert to /etc/network/interfaces? [09:31] TvL2386: could you use some other implementation to get the same effect? [09:31] I am normally a Debian weenie who happens to be fighting an Ubuntu server today, so I dunno [09:32] hehehe :) [09:32] But about 10 years ago, the important parts were "ifupdown" package and a udev rule to call net.agent [09:32] Also, systemd-networkd or network-manager can (partly) replace interfaces(5), so those might be options for you [09:33] It's just a bit frustrating that all these things worked perfectly in 16.04 and it seems like netplan isn't mature yet [09:33] yet /etc/network/interfaces is deprecated and not working anymore [09:33] I 100% commiserate; I had similar grief with upstart in 10.04 [09:34] yeah as per your bugs.debian link: netplan generates backend files in /run and hands it of the the network daemon. Which seems to be networkd on 18.04 [09:35] so I gotta check if I can make my own persistent "backend files" [09:35] oh! [09:35] TvL2386: if you run "networkctl" does it say unmanaged? [09:36] for this particular test I created a shell script to manually conffigure stuff via root cron using @reboot :) [09:36] checking [09:36] If netplan is writing systemd-networkd config in /run, then the answer is simply that you can write your own systemd-networkd config in /etc/systemd/network [09:36] I don't know if that supports dummyN but it's a place to start [09:37] it only says "unanaged" for interfaces not refered in my /etc/netplan/01-netcfg.yaml [09:37] ha :) [09:37] it does [09:37] yay [09:37] thx twb, that gives me some nice pointers to continue [09:37] the systemd index is "man systemd.directives" [09:38] man systemd.netdev and systemd.network should have all the relevant parts [09:38] You can also ask #systemd (must be registered nick). If systemd actually supports what you want to do, they will help. If systemd doesn't do it, they'll try to trick you. [09:39] nice, I see now what it does! My yaml config is used in /run/systemd/network/10-netplan-ens1f0.network [09:39] among other nics [09:39] hahaha twb [09:39] thx trippeh_ [09:40] SUPPORTED NETDEV KINDS [09:40] âdummy â A dummy device drops all packets sent to it. â [09:40] `man systemd.netdev` [09:40] cool [09:40] Table 1. Supported kinds of virtual network devices [09:40] long list :D [09:40] what's the use case for that? [09:40] vxlan... nice :) [09:40] as opposed to just rp_filter and blackhole routes? [09:41] well... if you wanna know: I am testing ECMP. Got a ubuntu 18.04 server which needs a loopback ip address (not ^127.*) [09:41] and I wanted to add it to a dummy nic [09:41] ip address add 4.3.2.1 dev lo [09:41] yeah yeah... [09:41] I want dummy! [09:41] ;) [09:42] fair enough, I'm just trying to understand why, for my own benefit [09:42] because I thought it was easy [09:42] haha [09:42] and the "lo" interface is magically configured [09:42] it has 127.0.0.1/8, though no netplan config [09:43] yeah I don't know where "ip address add 172.18.77.1/22 dev lo brd +" would go in systemd-networkd [09:43] the only obvious difference is that if you did that you'd get back ICMP responses instead of just nothing [09:43] I get that line until the "brd +" [09:43] broadcast +? [09:43] brd + just yep [09:43] what does that do? [09:44] hmmm... `grep -r 127.0.0.1 /run` does not return any hits [09:44] probably somewhere else of course [09:44] http://ix.io/1iWJ [09:45] if you don't brd +, there's no broadcast address set. Only matters if you're doing broadcasty things (e.g. mdns, I guess) [09:45] The other nice one is "ip address add en0 192.168.1.2 peer 192.168.1.1" if you're e.g. talking to a fresh router over a direct cable [09:46] weeeeeeeeeeiiiiiiiiiirrrrdddddd [09:46] #netfilter can talk to you about that stuff if you care [09:47] man ip-address : interesting that "peer" [09:47] it's a plain point-to-point so you don't need any /30 crap, like we did back in the dialup days [09:47] yeah it reminds me of pppoe [09:48] alrighty... back to dummy stuff [09:49] or finding out how I can add ip addresses to interface lo using netplan [09:49] without messing with the magical 127.0.0.1/8 [09:52] TvL2386: I expect you just want to write something like printf '[Match]\nName=dummy0\n[Network]\nDHCP=yes\n' >/etc/systemd/network/fnord.network [09:56] curl https://transfer.sh/gkGih/dummy.network [09:56] something like that [09:57] just need to find out how to "start" it [09:57] did `systemctl daemon-reload` [09:57] does daemon-reload work for networkd? [09:58] I thought that only reloaded /etc/systemd/system [09:58] I have no idea [09:58] it was a reflex :) [09:58] I would try restarting systemd-networkd and then look at networkctl [09:58] or, read the manpage :-) [09:59] yeah reading at the moment [10:03] I'm usually lazy and not in prod at that point so I just reboot the whole host [10:05] yeah same here... but I feel it's in my best interest to know how this works now that /etc/network/interfaces is deprecated [10:06] I made a dummy.netdev file that should generate the device [10:06] I made a dummy.network file that configures it [10:06] I'm looking for a way to reload networkd(?) so it "sees" this new configuration [10:07] systemctl restart systemd-networkd [10:08] yeah baby! [10:08] # networkctl list dummy0 [10:08] IDX LINK TYPE OPERATIONAL SETUP [10:08] 18 dummy0 ether routable configured [10:08] ip a s dummy0 # looks good to [10:08] cool [10:09] step 2: apt-get purge netplan [10:11] nice [10:11] I did that to NM for the first like 4 years after it landed, because it routinely broke my *wired* connections on servers [10:21] I'm just wondering if there's a more graceful way to alter the running configuration without completely restarting systemd-networkd [10:21] TvL2386: no need to purge (as that also removes some metapackages), just remove any config from /etc/netplan and it won't interfere. [10:22] true blackflow [10:22] The following packages will be REMOVED: [10:22] netplan.io nplan ubuntu-minimal [10:22] I assumed they wanted to remove it for the cathartic pleasure rather than any real need :-) [10:22] :) [10:22] wrt reconfiguring networkd without restarting, there's this (still open): https://github.com/systemd/systemd/issues/6654 [10:23] like when you get an old piece annoying of kit and smash it with a hammer [10:23] yeah but ubuntu-minimal.... my OCD would ahve an issue with removing that :) [10:24] hehehe :) [10:24] I don't really care tbh [10:24] Don't get me wrong, I care about your OCD, just not about ubuntu-minimal :P [10:24] ;) [10:26] netplan sounds nice on paper but in practice... I've hissed and barked at it here so it's not my intention to do that again. Gave it a chance and after several months no I still see no purpose of it. Yeah, centralized config regardless of backend, but in my experience any abstraction (and this is abstracion) is bound to either: a) do a half-assed job, or b) become extremely complex and thus [10:27] buggy, in order to satisfy all the functions of supported backends. [10:27] currently netplan is in the stage a) as it doesn't cover all the functions. [10:31] so `netplan apply` generates some .network files in /run/systemd/network and then restarts systemd-networkd [10:31] that ~3sec disruption I have on applying seems to be the same as when manually restarting systemd-networkd [10:32] I agree blackflow [10:32] that's how I experience it as well [10:33] why use netplan on ubuntu-18.04 if you can generate those .network files yourself. Have more control, less magic.... [10:33] same reason why I don't use ufw... [10:33] which btw is against systemd policy for generators, that are only supposed to generate unit files and symlinks. and going against systemd policy is bound to introduce breakage in the future, as sd developers don't like to care about stuff they recommend against. [10:34] enough whining from my side though :) [10:34] less is more. sometimes literally (via symlinks) :))) [10:54] cpaelzer: ok so whats the purpose of lcore and pmd threads in the context of dpdk? [11:10] jamespage: cpus the threads will spin on [11:11] PMD threads are the polling mode device drivers [11:11] you want those close to the device in numa systems [11:12] lcore you can think of the management plane a bit [11:13] like allocation of memory, some extra tasks [11:13] essentially all nonPMD work it does [11:13] masks for those can be set, but you really really have to know your HW to do so correctly [11:13] (cpu masks) [11:14] TL;DR: lcore = DPDK-EAL-thread; pmd-thread = a thread of the poll mode drivers(s) [11:14] jamespage: does that explain what you needed? [11:15] cpaelzer: yes - I think the charm is not quite doing the right thing at the moment [11:15] jamespage: I always liked the blog kevin wrote https://developers.redhat.com/blog/2017/06/28/ovs-dpdk-parameters-dealing-with-multi-numa/ [11:15] we set a lcore mask based on the number of cores to allocate basedon the numa topoligy [11:15] maybe that can help to get things straight [12:13] good morning [12:13] cpaelzer: I think I finally got hit by the glibc pending migration. DEP8 dependencies aren't installing anymore :/ [12:14] ahasenack: do you know how to resolce or should I show you in a quick session? [12:14] cpaelzer: this is in a bileto ticket for now [12:15] I tihnk we had that - where you called the interface ugly [12:15] cpaelzer: is the solution that horrible url mangling? :) [12:15] well you can do the same in the bileto tests [12:15] "interface" is a compliment :) [12:15] ahasenack: yes [12:15] https://bileto.ubuntu.com/excuses/3399/cosmic.html [12:15] alles kaput [12:15] locally I had to enable proposed (--apt-pocket=proposed) [12:15] at least then it worked [12:16] ahasenack: well that is much easier than selective unmasking then [12:16] you can to &all_proposed=1 (or it is a - instead of a _) [12:16] just harder to show, in the context of a merge proposal [12:16] ah, that [12:17] hm, there is no "retry" icon for sssd, just freeipa, why us that? [12:18] link? [12:18] just above [12:18] https://bileto.ubuntu.com/excuses/3399/cosmic.html [12:18] found it [12:19] because always failed = ok [12:19] well, that puts me in an odd situation [12:19] and retry is only shown in case it is not ok [12:19] since there were no dep8 tests before [12:19] I'm adding them for the first time [12:19] once they were successful once in the archive it will reset to expect that [12:19] and there is no way to tell bileto to use cosmic-proposed for the first run? [12:20] umm [12:20] you might still be able to retrigger it [12:20] "target series" should include proposed [12:20] in the list, I mean [12:20] I'm just nt sure it would pick up the updated result [12:22] hmm, no you can't reset it ... :-/ [12:22] hmpf [12:22] "You submitted an invalid request: Package sssd does not have any test results" [12:22] bileto has the chance to be so much more [12:22] is testing locally ok? [12:22] yes [12:22] I'm even testing a login [12:22] but I want to add more [12:23] cpaelzer: https://git.launchpad.net/~ahasenack/ubuntu/+source/sssd/tree/debian/tests/all-ldap?h=sssd-dep8-tests [12:26] smb: FYI iproute2 resolved [12:27] cpaelzer, yay! [12:27] 3/5 is already a much better rate [12:27] * cpaelzer feels bad to feel good about 3/5 success rate ?! [12:28] xnox: I restarted dbus once more, that was the only one failing (timeouts, nothing critical) [12:39] TvL2386: blackflow: because it's a) not magic, and b) not doing anything but writing unit files (with some small exceptions for NM+wpa, but hey) and c) some people don't know the whole syntax of systemd-networkd files, whether they should use a .network, .link, .netdev, or even a .rules file to do what they want to do. [13:02] cpaelzer: https://code.launchpad.net/~ahasenack/ubuntu/+source/snapper/+git/snapper/+merge/354142 ubuntu/devel hasn't moved after the new debian/sid release was imported [13:02] cpaelzer: I guess it's a bug when dealing with sync'ed packages [13:05] yes [13:05] didn't you file a bug last week? [13:05] when it came up the first time [13:12] jamespage: you might help me on an ambiguity as well [13:12] I can separate bweteen OVS and ovsdb [13:12] but those two /lib/systemd/system/openvswitch-switch.service /lib/systemd/system/ovs-vswitchd.service [13:13] it seems in the past I used the former but now have to use the latter [13:13] I did some work for bionic to switch to a more native aapproach [13:13] now you have ovsdb-server and ovs-vswitchd which are both part of openvswitch-switch [13:13] so restarting openvswitch-switch on old or new will dtrt [13:14] cpaelzer: ^^ [13:14] ok'ish [13:14] thanks [13:14] that explains what happened but now I'm back at my error [13:14] :-) [13:16] jamespage: now knowing this let me try to re-catch my issue [13:22] coreycb: bah we have an upgrade issue on the nova-common/python-nova twiddles from queens to rocky [13:25] ahasenack: reading your MP update - should I sponsor snapper? [13:25] it really LGTM [13:25] cpaelzer: yes please [13:25] and I see you can't [13:25] thought MOTU would be wih you already [13:25] I also tried it on my arm "box", worked just fine [13:25] isn't that MOTU [13:25] even though that one isn't armhf, it's armvl7 or something [13:25] I'm not motu :( [13:25] I'm.... andreas! [13:25] grml, you should be all of that by now [13:25] I'm trying [13:25] the dmb agenda is out of date [13:25] I just pinged ubuntu-devel about it [13:26] whatever it is, edit yourself in - maybe add a section "since below is out of date extra topics: -..." [13:26] I was punted by being shy as well, it was out of date and when it was back three other topics got in front of me [13:27] ahasenack: as a verification - a0e4e65631d63fe4fcf6b5938b2dc649f5f2a00f ? [13:27] ok [13:27] let me check [13:27] hm, no [13:27] let me see what is in lp [13:28] I might have pushed from my container [13:28] checking [13:28] a0e4e65 is on your remote [13:28] at least from my POV [13:29] but good that we check, let me know what you find [13:29] sure [13:29] jamespage: ok need a hand? [13:29] mine has debian/sid updated pointing at 0.5.6-2, the lp mp is against ubuntu/devel [13:29] jamespage: was just trying to figure out the horizon build failure in cosmic [13:30] c'mon git pull [13:30] * ahasenack waits [13:30] jamespage: kind of odd LP doesn't show the build log [13:30] coreycb: some other failure hit retry [13:30] jamespage: ok [13:33] cpaelzer: ah, I had updated dep3's last-update [13:33] coreycb: missing breaks/replaces [13:34] jamespage: ah, that'll do it [13:34] will fix and ping you [13:36] ok [13:39] When SYSTEM is running. Ethernet card is frozen, But the system still running. Connected to monitor with hdmi, ifup shows the card is up. lshw -c network shows the enp is correct. route -n shows the routes are okey. But why does the ethernet card then not respond? [13:39] Running 18.04 Server LTS [13:39] jamespage: neutron did not get an update for python(3)-neutron -> neutron-common. think i should do that now and consolidate *.install into neutron-common.install? [13:40] what do you mean the ethernet card is frozen ? [13:40] How do you mean the ethernet card does not respond. [13:40] Ussat: It SUDDENTLY stop responding [13:41] after some days running [13:41] OK, assume I have no idea what youre talking about, responds HOW ? what do you expect....how is it not "responding" [13:41] its a onboard ethernet card [13:42] Loose its IP ? can ping ? I am confused here [13:42] I cant go on internet, cant access it localy on my network [13:42] not losen ip. i cant ping. [13:42] it looks the same way as when it was suppose to work propperly [13:42] example: when i try to ping my gateway (Fibermodem) i get Destination host unreachable [13:43] When i try to ping my server inside/outside the network i get the same message from my laptop connected to my network [13:43] do you still have a default route ? [13:43] yes [13:44] ifconfig still shows IP etc ? [13:44] yes [13:44] this is a laptop ? [13:45] No. its not a laptop. its a high tower desktop computer set as server [13:46] hmm...well, have a meeting in 10 mins need to go to....will be back in a hr...sorry [13:46] its a MSI Z270 Gaming m3 (MS-7A62) mainboard [13:49] https://ubuntuforums.org/showthread.php?t=2381674 [13:58] cpaelzer: a0e4e65631d63fe4fcf6b5938b2dc649f5f2a00f is correct, my local change is irrelevant [14:02] ok ahasenack [14:03] also the debdiff looks equal to the last MP I reviewed [14:04] ahasenack: uploaded [14:04] thx [14:19] coreycb: ok but make sure you pull before you do - just uploaded a fix for the metadata/ssl/san/ipaddress cells v2 issue gnuoy has [14:20] cpaelzer: bearing in mind you have to enable dpdk in later ovs versions, do we need to have a separate -dpdk package any longer? [14:20] or can we just bake this into the standard binaries [14:20] jamespage: ok [14:21] jamespage: you wanted to keep it split to avoid regressions into "normal" OVS [14:21] but ack, given that we enable it it is much safer today [14:21] cpaelzer: earlier versions enabled dpdk blindly [14:21] you might remove it in an early 19.04 upload maybe [14:22]