[00:46] <Glorfindel> uhhhhhhhhhhhhghghghg. so. I created a  xfs filesystem because I ran out of inodes on ext4, but I've also filled the inodes on this xfs disk. is there any other filesystem I could use that has more inodes?
[00:48] <sarnold> woah, really?? I didn't realize that was possible with xfs. I thought it was dynamic. heh.
[00:50] <Glorfindel> wait.... ncdu only reported 16gb used, but df -h shows I used all 20gb?
[00:50] <Glorfindel> maybe I didn't run out inodes, that would be good
[00:50] <sarnold> that makes more sense to me :)
[00:50] <Glorfindel> at the same time though, over 20gb of data to create a tileserver for a map that's kept in a ~200mb database seems a tad excessive
[00:51] <Glorfindel> 7.9 million inodes.... yeah I'm submitting a bug that's ridiculous
[00:55] <sarnold> I suspect that means whoever made the database design really understood the principles of data normalization! :D
[00:56] <Glorfindel> lol
[00:56] <RoyK> and whoever wrote that software, didn't understand that millions of files is generally a bad idea…
[00:57] <RoyK> Glorfindel: keep in mind that metadata also takes up space
[00:58] <sarnold> Filesystem                       Inodes    IUsed       IFree IUse% Mounted on
[00:58] <sarnold> srv                         12875952530       10 12875952520    1% /srv
[00:59] <sarnold> oh. don't mind me. I can't read.
[00:59] <RoyK> 12 billion - should be enough for everyone (tm)
[00:59] <sarnold> :)
[00:59] <Glorfindel> one would think :o
[01:00] <sarnold> 62 million files on that thing, give or take.
[01:00] <sarnold> suddenly 8 million for a mineserver map server seems pretty silly
[01:00] <RoyK> or someone aught to think that 8 million files in a directory (or tree) just means "you nee to use a database!"
[01:01] <RoyK> *need*
[01:01] <Glorfindel> sarnold: wow
[01:02] <sarnold> Glorfindel: my ubuntu mirror is ~800k inodes.. the unpacked sources are way bigger :)
[01:05] <Glorfindel> heh, makes sense
[02:29] <keithzg[m]> Huh, I wonder why multiple VMs of mine (three so far) are running into kernel panics on boot with the latest kernel :/
[02:35] <sarnold> :(
[02:36] <RoyK> keithzg[m]: which kernel is that?
[02:37] <keithzg[m]> Specifically, I'm seeing `Kernel panic - not syncing: VFS: Unable to moutn root fs on unknown-block(0,0)`. Kernel is "4.15.0-38-generic #41-Ubuntu SMP Wed Oct 10 10:58:23 UTC 2018 i686 athlon i686 GNU/Linux"
[02:38] <mason> keithzg[m]: Ah, sounds like your initramfs files didn't build...?
[02:38] <mason> keithzg[m]: Can you boot into older kernels and rebuild your initramfses?
[02:38] <RoyK> update-initramfs -a
[02:38] <RoyK> perhaps
[02:38] <keithzg[m]> mason: Yeah that's what I thought, although `sudo update-initramfs -u -k 4.15.0-38-generic` did not fix it.
[02:39] <keithzg[m]> RoyK: "Illegal option -a"
[02:39] <mason> -c -k all, isn't it?
[02:39] <mason> I confuse platforms sometimes.
[02:39] <mason> Yeah, try update-initramfs -c -k all
[02:40] <keithzg[m]> Shall do...
[02:42] <RoyK> keithzg[m]: out of curiousity - what sort of virtualisation?
[02:42] <keithzg[m]> I wonder if this all has to do with grub2, I got the "GRUB upgrade scripts have detected a GRUB Legacy setup in /boot/grub" message last week and accepted chainloading, intending on actually checking things sometime this week. All the VMs in question are quite old ones I inherited from my predecessor, they started out on at least 10.04!
[02:42] <keithzg[m]> RoyK: KVM
[02:42] <RoyK> ok
[02:43] <RoyK> 10.04 is a wee bit old ;)
[02:43] <keithzg[m]> Yup they've had long lifespans, gone to 12.04 then 14.04 then 16.04 and recently finally 18.04 and a move to an entirely different machine a few months back!
[02:44] <keithzg[m]> Alas, `update-initramfs -c -k all` did nothing, the VM I tried still kernel panics.
[02:44] <mason> keithzg[m]: Hrm. I'd want to rip apart the initramfs at this point. Unsure what else to do.
[02:45] <keithzg[m]> Very oddly, the VM I first noticed this on earlier today worked fine after I rebooted into the prior kernel, applied pending updates, and rebooted again. It's also one of the old VMs.
[02:45] <mason> keithzg[m]: Make sure your /boot isn't out of space, although I'd think you'd see an error message to that effect when building.
[02:46] <mason> keithzg[m]: Another less-intense thing might be to delete the new kernel and reinstall, and capture the session in script(1) for further debugging.
[02:47] <RoyK> if it's out of space, remove old kernels - you may have to truncate some of the files to remove before running apt remove/purge, since apt uses some disk space just to remove things
[02:47] <RoyK> don't remove them manually - apt will be upset
[02:47] <RoyK> just run something like
[02:47] <RoyK> > /boot/somekernel
[02:48] <RoyK> to truncate it
[02:48] <keithzg[m]> mason: Yeah, these VMs don't even actually have a separate /boot partition, and / has plenty of space. Funny enough, the VM that recovered *had* run out of space, and I had initially thought that was a problem. Maybe in fact it was a problem on *that* VM and the same panic message is actually from a different underlying cause on the other VMs, although that seems like a *huge* coincidence if so.
[02:50] <mason> keithzg[m]: The initramfs having issues or being invalid is the most common cause of that error message. Kernel tries to load it and fails.
[02:51] <mason> keithzg[m]: Might be worth update-grub as well, with a spelunk into the grub config files to make sure it's specifying the right files.
[02:51] <keithzg[m]> mason: yeah already tried `update-grub`.
[02:53] <keithzg[m]> Oho, I fixed it on one of the VMs at least; under the presumption that there's some sort of reason why last week that I was still on grub-legacy was flagged during the `apt upgrade` I ran on 'em all, I tried running the suggested `upgrade-from-grub-legacy` and then rebooted. Et voila!
[02:53] <keithzg[m]> Thank goodness for virtualization, if I hadn't been able to take a quick snapshot before hand it might have been quite a while before I worked up the courage to try that ;)
[02:54] <mason> Ah, good.
[02:58] <keithzg[m]> Other failing VMs seem to have been fixed by the same approach. Other than "dpkg-maintscript-helper: error: environment variable DPKG_MAINTSCRIPT_NAME is required" as the last output line, `update-from-grub-legacy` seems to have worked fine and resolved this. Phew!
[04:02] <n00bee> I upgraded a VPS from 16.04 to 18.04 and now there is no network. The VPS lets me log in via a web console but I can't even ping out from the server. How to troubleshoot?
[04:03] <n00bee> The outputs for some of the commands are here: https://imgur.com/a/cmqGEfE (sorry it's an image rather than text paste but the webconsole doesn't allow text copy)
[04:05] <nacc> n00bee: try `sudo systemctl status networking` and if it's off, try `sudo systemctl start networking` ?
[04:06] <nacc> n00bee: your interface has no IP address
[04:07] <n00bee> nacc: It's probably worse. It says Unit networking.service not found
[04:07] <n00bee> *could not be found
[04:07] <nacc> i might have the wrong service
[04:09] <nacc> n00bee: try `sudo ifup eth0` ?
[04:09] <n00bee> nacc: sudo  systemctl restart NetworkManager.service ... also not found
[04:10] <n00bee> nacc: ifup command not found
[04:11] <nacc> n00bee: you're not on a desktop, so no NM
[04:11] <nacc> n00bee: ah it would appear that ifupdown maybe got purged
[04:12] <nacc> n00bee: you have two options, i think, follow what is written here: https://wiki.ubuntu.com/BionicBeaver/ReleaseNotes#New_since_16.04_LTS
[04:12] <nacc> n00bee: apt-cache policy ifupdown?
[04:13] <n00bee> nacc: apt-cache policy ifupdown suggests it may be installed
[04:13] <n00bee> wait..sorry..no
[04:13] <n00bee> installed: (none)
[04:14] <nacc> systemctl status systemd-networkd
[04:14] <n00bee> candidate: 0.8.17ubuntu1.1
[04:14] <nacc> may give you some debugging output
[04:14] <nacc> yeah, so if you want to keep using /e/n/i you need to have that installed, read the release notes above
[04:15] <n00bee> nacc: Sorry, what does /e/n/i mean?
[04:15] <nacc> n00bee: /etc/network/interfaces (which is what ifupown parses)
[04:15] <nacc> n00bee: sorry, gotta step away, read those docs
[04:15] <n00bee> nacc: ok. thanks
[04:55] <n00bee> nacc: So I read the docs and they specify the 18.04 counterparts to 16.04 commands and services. Using that I was able to start systemd-networkd and do the equivalent of ifup eth0. But still no network.
[04:56] <d-rock> When you say "no network", what do you mean?
[04:56] <d-rock> Does ifconfig show eth0 up and running?
[04:56] <n00bee> d-rock: I upgraded from 16.04 to 18.04 today and now have not network. Can't SSH in or ping out.
[04:57] <d-rock> OK, but "ifconfig eth0" shows that the interface is UP and RUNNING, and shows tx/rx activity?
[04:57] <n00bee> d-rock: The web console doesn't allow copy paste. Let me take a screenshot of the output of ifconfig eth0
[04:58] <genewitch> looking for a package (or whatever) that is like pfsense, a web frontend for NAT/router/gateway/firewall.. the simpler the better. I literally just need NAT/dhcpd and i will be running pihole for DNS
[04:58] <d-rock> I'm on a terminal, so I can't view graphics :)
[04:58] <n00bee> d-rock: :) so it says flags=4098 broadcast, multicast
[04:58] <d-rock> That's not up
[04:59] <n00bee> d-rock but it doesn't say anythign that indicates up or running. All packets RX, TX are 0
[04:59] <genewitch> also, i have eth1, but it's not getting a static IP, i have /etc/network/interfaces.d/eth1 populated
[04:59] <d-rock> For reference, my interface: enp5s0: flags=4163<UP,BROADCAST,RUNNING,MULTICAST>  mtu 1500
[04:59] <d-rock> First, let's see if we can manually configure it
[04:59] <d-rock> Unless you've already verified that
[05:00] <genewitch> http://projectftm.com/#pAGR1e0AkCvAlV8xVaYeAA this is the eth1 file
[05:00] <n00bee> d-rock: OK wait. It is up now. I ran sudo ip link set eth0 up
[05:01] <n00bee> after that ifconfig eth0 shows up, broadcast, running, multicast
[05:01] <d-rock> OK, does it have an IP set?
[05:01] <d-rock> I suspect you'll need to re-ifup
[05:01] <d-rock> I can't remember the networkd command off the top of my head
[05:02] <n00bee> d-rock I don't know if it has an IP set. ifconfig does have a line for inet6 but ipv6 is not enabled for the server i think
[05:02] <d-rock> inet6 will always be set. At the very least, it gets the link-local address
[05:02] <d-rock> Do you see a line for just plain "inet"?
[05:03] <JanC> RoyK: a filesystem *is* a database
[05:03] <n00bee> d-rock: No line for inet
[05:03]  * JanC doesn't understand why people don't understand that
[05:03] <d-rock> OK, can you re-run the networkd command?
[05:04] <n00bee> d-rock: you mean sudo systemctl restart systemd-networkd
[05:04] <d-rock> That'll do it
[05:04] <d-rock> There's another, more surgical command, but I can't remember it
[05:05] <n00bee> d-rock: That didn't do anything :(
[05:05] <n00bee> It didn't throw any errors but I still can't ping out. And ifconfig eth0 still doesn't show an inet
[05:05] <d-rock> Then let's just try to bring this up manually
[05:06] <d-rock> ip addr add <ip/mask> dev eth0
[05:06] <buddhirt> some query if anyone knows? i have installed isc-dhcp-server in 18.04. Seems like in 18.04 interface is not auto-up without connecting network cable to client machine, so dhcp server cannot listen to defined interface. Any solution
[05:07] <n00bee> d-rock: I have the static IP address. and in /etc/network/interfaces the netmast is 255.255.255.0. So you're saying the sommand is ip addr add 192.241.x.x/255.255.255.0 ?
[05:07] <genewitch> are you freaking kidding me
[05:07] <genewitch> network configuguration is done through YAML?
[05:07] <genewitch> come on, guys, seriously
[05:07] <genewitch> ubuntu drives me batty
[05:07] <d-rock> genewitch: I kinda had the same reaction when I saw netplan
[05:08] <d-rock> n00bee: ip addr add 192.241.x.x/24 dev eth0
[05:08] <n00bee> d-rock: Thanks!
[05:08] <d-rock> 255.255.255.0 is a 24 bit mask
[05:09] <n00bee> d-rock: OK so now I see that in ifconfig eth0. Now restart systemd-networkd ?
[05:09] <d-rock> No, let's hold off
[05:09] <d-rock> Can you ping your gateway now?
[05:09] <JanC> you don't have to use netplan
[05:10] <n00bee> d-rock: Web console hung
[05:11] <d-rock> n00bee: when you set the ip address?
[05:11] <n00bee> d-rock: yay! i can ping the gateway
[05:11] <d-rock> Huzzah!
[05:11] <d-rock> ip route add 0/0 via <gateway>
[05:12] <n00bee> d-rock : done
[05:12] <d-rock> Should be able to ping, say, 8.8.8.8
[05:12] <n00bee> YES!!
[05:12] <JanC> and netplan makes lots of sense when you have to do cloud configurations
[05:12] <d-rock> Should be able to reach it from outside, as well
[05:13] <n00bee> d-rock: yes, it's back online and accessible from the outside as well
[05:13] <n00bee> d-rock: are these changes going to stick after a reboot?
[05:13] <d-rock> They will not
[05:14] <d-rock> But, it does confirm that this is an issue with networkd and/or netplan, not the NIC or network itself
[05:14] <n00bee> d-rock: hmm...so what should i do
[05:14] <JanC> or with the netplan configuration ;)
[05:15] <d-rock> Sure, it could be the config
[05:15] <d-rock> To be honest, I'm fighting my own battle with 18.04 networking. I ended up just writing a shell script to init things the way I wanted
[05:15] <JanC> there is a #netplan channel to discuss that BTW
[05:16] <n00bee> d-rock: I could probably just move stuff out of this server and set up a new one from scratch
[05:16] <d-rock> I hesitate to project my own experience, but that might be simpler :P
[05:17] <genewitch> so is there a simple answer? webgui for all the firewall-y stuff in linux? Like pfsense/monowal/ipcop for BSD
[05:17] <d-rock> In any case, I need to drop off. Good luck!
[05:18] <n00bee> d-rock: Thanks much. You were a life saver and hero today!
[05:18] <d-rock> NP, glad I could repay some of the help I've gotten on this channel :)
[05:18] <JanC> (although e.g. cyphermox is in here too it's easier for them to miss discussions of netplan maybe)
[05:20] <genewitch> is there a guide to having 3 routable networks under netplan?
[05:20] <genewitch> i have 10.0.0.0/8 192.168.1.1/24 and WAN
[05:20] <genewitch> which is technicaly 192.168.42.1/something
[05:23] <JanC> genewitch: that's exactly the sort of thing you want to ask in the #netplan channel (if it's not in their documentation)  ;)
[05:23] <genewitch> GAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAH
[05:24] <JanC> and do you really need a web interface for the firewall? or is a descriptive configuration file also useful?
[05:24] <genewitch> ubuntu implemented netplan
[05:24] <genewitch> they broke 10 years of working network config
[05:24] <JanC> you don't have to use netplan if you don't want to...
[05:24] <JanC> ifupdown still works if you prefer that
[05:24] <genewitch> /etc/network/interfaces.d/eth1
[05:24] <genewitch> doesn't work
[05:25] <genewitch> this used to be stupidly obnoxiously straightforward, it's taken hours out of my night
[05:26] <JanC> if you don't have a netplan config and ifupdown is installed, that should still work, I guess?
[05:26] <genewitch> i have a netplan config but it is complaining about something
[05:26] <JanC> if not, file a bug report   :)
[05:27] <katamo> genewitch you should be able to `netplan generate --debug` and get more verbose information on what it errors about
[05:28] <genewitch> http://projectftm.com/#EfzbyVArZW8ykjaekYCE_w
[05:28] <genewitch> exact same output
[05:29] <genewitch> Error in network definition //etc/netplan/eth1.yaml line 5 column 17: expected sequence
[05:32] <JanC> shouldn't addresses take a sequence?
[05:33] <JanC> simply said: put it between []
[05:34] <JanC> (I'm not a netplan expert, so this is just a guess, to be fair)
[05:34]  * katamo had a toddler interruption
[05:34] <genewitch> yeah that's all working now, but the routes are wrong
[05:36] <JanC> you have "addresses: [10.0.0.1]" instead of "addresses: 10.0.0.1" now?
[05:37] <katamo> http://projectftm.com/#FD5qmugC0YFFuZUHIlvuFQ
[05:37] <katamo> genewitch ^
[05:37] <katamo> does that get closer?
[05:38] <katamo> note the use of space between addresses & eliminating the genmask/netmask lines in favor of */* address syntax
[05:38] <genewitch> JanC: yes
[05:38] <genewitch> i had that originally :-) but it doesn't look right, you know?
[05:38] <genewitch> a single IP is a /32
[05:40] <genewitch> katamo: i did addresses: [10.0.0.1/9]
[05:40] <genewitch> er /8 obviously
[05:41] <katamo> genewitch I just ran `netplan generate --debug` after copy paste of the link I shared with you. it completed without error
[05:42] <genewitch> katamo: yes mine works fine too, except i can't ping 10.0.0.2 :-)
[05:42] <katamo> `ip r | pastebinit`
[05:43] <genewitch> i can't do that, because the internet isn't working now
[05:44] <katamo> fair enough lol. whats the output?
[05:44] <genewitch> http://projectftm.com/#n3T06ci4QHs35bHBQYdPWQ
[05:44] <genewitch> lemme if down eth1 this is ridic
[05:44] <katamo> `ip r del default via 192.168.1.1 dev eth0`
[05:44] <katamo> for testing
[05:45] <katamo> or that works
[05:45] <genewitch> i can ping 192.168.1.1 after running that, but nothing else
[05:46] <genewitch> default route is for what, internet?
[05:46] <JanC> default route is for everything that doesn't have a specific route
[05:47] <JanC> usually that means the general internet  ;)
[05:48] <katamo> Okay, can we get the full output for configs on all interfaces?
[05:48] <katamo> for one thing in the "routes", your 192.168.1.1 dev eth0 network has no netmask definition
[05:48] <genewitch> http://projectftm.com/#ZbbJinc8BCQs-wI-SOKPuQ
[05:49] <genewitch> 192.168.1.0/24 is going to be out of band, the only two networks that should be routable outbound are 10.0.0.0 and whatever usb0 is
[05:49] <katamo> config files. either netplan or interfaces(.d)/* ?  sorry I'm not used to reading ifconfig output any more lol. woe is me
[05:49] <genewitch> the yaml?
[05:50] <genewitch> eth1.yaml http://projectftm.com/#HKWqhJ0ub1YeoPaZs72tUQ
[05:50] <genewitch> eth0.yaml http://projectftm.com/#3fWxIdIZ-5dpFLaSRUr-WQ
[05:50] <genewitch> there's no usb0 in either location
[05:51] <genewitch> usb0 is brought up with dhclient usb0
[05:52] <genewitch> i wanna make sure i can still access the internet if i disable eth1.service
[05:52] <JanC> the output of "ip route" might be relevant...
[05:52] <katamo> okay. lets do one interface at a time then. that was a good idea. and i'm still hung up on "192.168.1.1 dev eth0"
[05:52] <katamo> are you okay with bringing all interfaces down, then raising eth1 first? sounds like you're concerned about that.
[05:53] <genewitch> ip route with `systemctl disable eth1` http://projectftm.com/#bTUtehwmWwosE68B7sj8_A
[05:53] <katamo> JanC I think we have that output here: http://projectftm.com/#n3T06ci4QHs35bHBQYdPWQ
[05:53] <genewitch> i can ping both public (4.2.2.2) and private (192.168.1.12) with that setup
[05:53] <JanC> (remember 'ifconfig' has been deprecated for over a decade and in theory might disappear any day now :P )
[05:53] <katamo> Okay, we should not have multiple default routes.
[05:54] <JanC> ah, right
[05:54] <genewitch> i want eth1 to provide NAT from usb0 to the 192.168.1.1 network
[05:54] <genewitch> but not via this device
[05:54] <genewitch> I got a router connected to eth1
[05:54] <katamo> we're doing routing from this box?
[05:54] <genewitch> on the WAN port
[05:54] <genewitch> NAT only
[05:55] <genewitch> well i guess routing, but i just want to be able to ping 10.0.0.2, 192.168.1.1, 192.168.1.12, 4.2.2.2, google.com
[05:55] <genewitch> once i am there, i know what to do
[05:56] <katamo> Okay. can we bring the eth0 down? `ip link set eth0 down` ?
[05:56] <genewitch> katamo: should i reboot with eth1 enabled first?
[05:56] <genewitch> and this is going to slow me down since i have to physically access the machine which is not in this room
[05:57] <katamo> wait a sec
[05:57] <genewitch> eth1 systemd service is disagbled right now
[05:57] <genewitch> and everything is working as i'd expect
[05:58] <genewitch> i know i am rough
[05:58] <katamo> okay you specifically mentioned every network except the 192.168.1.0/24 dev eth0 network
[05:58] <katamo> what is the .0/24 network?
[05:59] <katamo> and why does .1.0/24 have a netmask/24 whereas the 1.1 network has implied /32?
[06:00] <genewitch> inet 192.168.1.6  netmask 255.255.255.0  broadcast 192.168.1.255
[06:00] <genewitch> 192.168.1.1/24? it's a class C handed out by another router right now
[06:00] <genewitch> it's out of band
[06:01] <katamo> admittedly i'm a server engineer not a networking engineer and I deal mostly in virtual networks/servers
[06:02] <genewitch> yeah this is a unique thing.
[06:02] <genewitch> It's a SBC with two nics that actually operate at gigabit so i was hoping i could set this up
[06:02] <katamo> sounds like a fairly complex network for usb nic's?
[06:02] <katamo> ah
[06:02] <genewitch> the usb nic is a cellphone
[06:02] <genewitch> my cellular modem got taken out by lightning
[06:02] <katamo> oic. cellular gateway?
[06:03] <genewitch> yes
[06:03] <genewitch> so usb0 is just going to NAT to eth1
[06:03] <katamo> I use LXD ubuntu servers as routers with odd nic devices & client wifi radios as wan all day so it should be straightforward
[06:03] <katamo> hrm
[06:03] <genewitch> eth0 is just out of band so i can get in if those are down
[06:04] <genewitch> As i said, i just need to ssh IN to eth0, eth0 doesn't need to route anywhere
[06:04] <katamo> Okay, do you ssh into it via the 192.168.1.1 or 192.168.42.129 address?
[06:05] <genewitch> 192.168.1.6
[06:05] <genewitch> .42 is cellular
[06:05] <genewitch> ip a without eth1 http://projectftm.com/#iXR_MMRn9J7DXpLxooHptQ
[06:06] <genewitch> all this routes. i can ping google.com and i can ssh in
[06:06] <genewitch> lemme reboot with eth1 enabled
[06:06] <genewitch> baby steps
[06:07] <katamo> Okay http://projectftm.com/#lmhaVjJrGSozjkZ5hIKL0A
[06:07] <katamo> that accurate?
[06:08] <genewitch> yes
[06:09] <genewitch> working networking http://projectftm.com/#loQFlNCyXBQso4LWCf84UA
[06:09] <genewitch> not working networking http://projectftm.com/#wnB3yyIcgi4U-81DDvd3SQ
[06:11] <katamo> okay, please amuse me for a sec. can you cp your eth1.yaml to a safe place, then make a new one with copy paste of http://projectftm.com/#OMMBbGahZEd0nkjU5hvALw ?
[06:11] <genewitch> of course
[06:11] <katamo> then `ip link set eth1 down; ip link set eth1 up`
[06:11] <katamo> wait
[06:12] <katamo> `ip link set eth1 down; netplan generate --debug && netplan apply && ip link set eth1 up`
[06:13] <genewitch> stand by
[06:14] <genewitch> ok
[06:14] <katamo> how's our ping 4.2.2.2 look?
[06:14] <genewitch> can't ping 4.2.2.2
[06:14] <katamo> :*(
[06:14] <genewitch> :-D
[06:15] <genewitch> From 10.0.0.1 icmp_seq=4 Destination Host Unreachable
[06:15] <genewitch> so it thinks 10.0.0.1 is the default route
[06:15] <genewitch> whereas it should be 192.168.42.whatever
[06:15] <katamo> if its not a route, than remove the gateway line from your .yaml
[06:15] <genewitch> ^
[06:15] <katamo> wait.... this is.... duh. its trying to use its own interface as the default route
[06:16] <katamo> shoulda seen that sooner. i'm silly after hours
[06:17] <genewitch> i'm just happy ssh always works on reboot right now
[06:18] <katamo> haha no doubt. you shouldnt have to reboot between netplan config changes though
[06:18] <genewitch> i do anyhow
[06:18] <genewitch> it's like 6 seconds
[06:18] <katamo> fair enough then
[06:18] <genewitch> okay now i can ping 192.168.1.12 (that network) and the internet
[06:19] <katamo> woohoo!
[06:19] <katamo> and everything else?
[06:19] <genewitch> so why is gateway not necessary? does that autoconfigure a route or something?
[06:20] <katamo> gateway tells the OS that it is a route. IE if there's not an obvious alternate, that is where to pass traffic for up stream nat
[06:20] <genewitch> everything else isn't set up yet and i imagine the reason i can't ping 10.0.0.2 is that the router that is connected to has "ignore WAN ping" rofl
[06:20] <katamo> likely
[06:20] <katamo> cell phones are a pita
[06:20] <genewitch> if this works it saves me literally $200
[06:20] <genewitch> so that's why i am so dedicated
[06:21] <katamo> absolutely! I've used my cell hotspot as WAN for 2 days for fail over
[06:21] <genewitch> that cellular modem costs a bit, and even that is double NAT - if i have to deal with double NAT i'd rather do it on my software
[06:21] <genewitch> oh this is my primary internet
[06:21] <katamo> :O
[06:22] <genewitch> i'm using my personal cellphone for testing, the actual cellphone is on a windows machine, i'm talking to you on an ubuntu VM in Los Angeles while sitting at a windows desktop in louisiana
[06:22] <katamo> do what works lol
[06:23] <genewitch> indeed.
[06:26] <katamo> genewitch given your use-case can I ping you with a thought out of channel? Its not ubuntu-server specific but relevant
[06:26] <genewitch> sure
[06:27] <katamo> ping sent
[12:19] <ahasenack> good morning
[12:46] <ahasenack> kstenerud: I still see two issues in d/changelog in https://code.launchpad.net/~kstenerud/ubuntu/+source/tmux/+git/tmux/+merge/357991, not sure if you are still working on that or waiting for another review pass
[12:55] <ahasenack> xnox: hey, question about a server seed change you made a while ago
[12:55] <xnox> ahasenack, Which one? ^_0
[12:56] <ahasenack> xnox: you added both gnupg and dirmngr to server seeds, but gnupg has a depends on dirmngr, is the latter really necessary?
[12:56]  * xnox does a lot of seed changes
[12:56] <ahasenack> xnox: https://pastebin.ubuntu.com/p/dnDZDwy8Md/
[12:56] <xnox> ahasenack, need to check things...... back in the day gnupg and dirmngr got split out and at one point did not have the dep between themselves.
[12:57] <ahasenack> could be
[12:57] <xnox> but now gnupg is a metapackage effectively
[12:57] <kstenerud> ahasenack: I've pushed changes to address that. Do I need to write a comment about it?
[12:58] <ahasenack> kstenerud: I don't know if just pushing changes sends out an email notification, a ping/comment would be nice
[12:58] <ahasenack> kstenerud: but I see two other issues there still, I don't think he commented on them
[12:58] <ahasenack> kstenerud: the bug number is missing a #,
[12:58] <ahasenack> kstenerud: and the new version number is incorrect
[12:58] <xnox> ahasenack, right. So i think the change still stands. Albeit redundant now, but was not much so before.
[12:59] <kstenerud> what should the version number be?
[12:59] <xnox> ahasenack, this could be made lighter, by replacing "gnupg dirmngr" with "gpg dirmngr gpg-agent"
[12:59] <ahasenack> kstenerud: check the security team's page with that table
[12:59] <xnox> ahasenack, but imho having the full gnupg suite of utils is nice.
[12:59] <ahasenack> xnox: I agree with having gpg and dirmngr for the reasons you stated, I was just wondering if I was missing something since gnupg has a depends on dirmngr
[12:59] <xnox> ahasenack, any particular reason why you are asking about this? do you want to drop all of gnupg? or does this gnupg-metapackage-of-doom pulling in too much now?
[13:00] <xnox> ahasenack, when we still had gnupg2 package and were transitioning to the new one, the dirmngr depends was not there.
[13:00] <xnox> so historical.
[13:00] <ahasenack> xnox: ok, thanks
[13:00] <xnox> but i'd want to keep it there, such that it doesn't regress again
[13:01] <ahasenack> agreed
[13:01] <xnox> cause to talk to keyservers one needs dirmngr installed.
[13:01] <ahasenack> yes, the network part
[14:03] <y1ds> Hello
[14:04] <y1ds> i`m having some problems with slapd on ubuntu 16.04 and was wondering if someone has some tips
[14:05] <y1ds> I want to change my admin password but I keep getting inufficient acces errors, something like this http://www.mehic.info/2014/05/rootdn-ldap_add-insufficient-access-50/
[14:05] <y1ds> but i dont want to go and change those files
[14:05] <y1ds> ive noticed a few thinfs
[14:06] <y1ds> in my olcDatabase={1}hdb,cn=config i have an olcRootDN and olcRootPW, this is the password i want to change
[14:06] <y1ds> but it seems i need to use the creds from olcDatabase={0}config
[14:06] <y1ds> but in there i only see a olcRootDN, no olcRootPW
[14:06] <y1ds> i`m reading here https://help.ubuntu.com/lts/serverguide/openldap-server.html
[14:07] <y1ds> and in the post install chapter i see this command: sudo ldapsearch -Q -LLL -Y EXTERNAL -H ldapi:/// -b cn=config dn
[14:07] <y1ds> this returns No such object (32)
[14:07] <y1ds> but i see the file strucute that is above in the docs
[14:07] <y1ds> any clues?
[14:09] <y1ds> ive tried ldappasswd -Y EXTERNAL -H ldapi:/// -s welkom123 cn=Manager,dc=yo,dc=lo
[14:09] <y1ds> ive tried making an ldiff and trying to put it in using my admin account but im not allowed
[14:31] <ahasenack> y1ds: do you know the existing password for cn=admin?
[14:33] <y1ds> well i know the password for manager, wich is the rootdn of the database, but not of the config
[14:35] <ahasenack> y1ds: try this example:
[14:35] <ahasenack> ldappasswd -x -D cn=admin,dc=lxd -w secret -s newsecret cn=admin,dc=lxd
[14:35] <ahasenack> -w: existing password
[14:35] <ahasenack> -s: new password
[14:35] <y1ds> that doesnt work
[14:35] <ahasenack> -D: dn of the admin
[14:35] <ahasenack> it does, I just did it
[14:35] <ahasenack> cn=admin,dc=lxd: dn whose password you want to change (same as -D for this casE)
[14:36] <ahasenack> note I provided the existing password
[14:36] <y1ds> yes i tried that
[14:36] <ahasenack> because the acl for the userPassword attribute reads, among other things: by self write
[14:36] <ahasenack> root@xenial-foo:~# ldappasswd -x -D cn=admin,dc=lxd -w secret -s newsecret cn=admin,dc=lxd
[14:36] <ahasenack> root@xenial-foo:~# echo $?
[14:36] <ahasenack> 0
[14:36] <ahasenack> I jsut did it
[14:37] <y1ds> sec
[14:37] <ahasenack> there are two admins here, btw
[14:37] <ahasenack> one is for the cn=config db, the other is for the dc=lxd (or whatever domain you have) db
[14:38] <ahasenack> my line above changes the pw for the dc=lxd suffix
[14:39] <y1ds> ldap_bind: Invalid credentials (49)
[14:39] <ahasenack> then -w has the wrong password
[14:39] <ahasenack> or -D is incorrect
[14:40] <y1ds> it is not
[14:40] <ahasenack> to change olcRootPW in the cn=config db (for the olcRootDN "user"), you need ldapmodify, not ldappasswd
[14:40] <ahasenack> ldappasswd only changes the userPassword attribute
[14:41] <ahasenack> try ldapwhoami with the same -D and -w
[14:41] <y1ds> ldap_bind: Invalid credentials (49)
[14:42] <y1ds> yeah i tried changint it with ldapmodify and an ldiff
[14:42] <y1ds> but i think the problem is that i need to be the root of olcDatabase={0}config
[14:43] <y1ds> there is also a olcRootDN in there, namely cn=config, but not a olcRootPW
[14:43] <y1ds> yeah what you say aboce
[14:43] <y1ds> above
[14:44] <ahasenack> it's not clear to me which password you want to update
[14:44] <y1ds> i think i need  to be the one for the cn-config db, but that one does not have a password
[14:44] <ahasenack> RootDN is the admin of the database, and it doesn't have to have an entry in the directory
[14:44] <ahasenack> its password is RootPW
[14:44] <ahasenack> that's like the old rootpw setting in the old slapd.conf
[14:44] <ahasenack> you need an ldapmodify operation to change that, authenticated with -Y EXTERNAL
[14:45] <y1ds> well, i want to change the password for the admin user of the dc=blabla, but i think i need the password for the admin of the cn=config db
[14:45] <ahasenack> you don't have the existing password corresponding to the userPassword attribute of cn=admin,dc=blabla?
[14:45] <y1ds> yes i do
[14:46] <ahasenack> then prove it with ldapwhoami, using that DN and password
[14:47] <y1ds> ah yeah that works
[14:47] <y1ds> made a typo before
[14:47] <ahasenack> in this example, https://pastebin.ubuntu.com/p/QjbPKCSkRZ/, you want to change userPassword from cn=admin,dc=lxd in line 24?
[14:48] <ahasenack> for the example above, it would be: ldappasswd -x -D cn=admin,dc=lxd -w <existingpassword> -s <newpass> cn=admin,dc=lxd
[14:49] <ahasenack> that treats cn=admin,dc=lxd just like any other entry, nothing special about it being admin, because you are binding as the entry itself
[14:49] <ahasenack> if you didn't have the existing password, then you would have to bind as the rootdn, and that password is rootpw, defined in the cn=config suffix
[14:49] <ahasenack> confusing, agreed
[14:49] <ahasenack> I'm not a super fan of this cn=config structure
[14:52] <y1ds> ah okay no but its differentin my setup
[14:52] <y1ds> the admin is not a user
[14:53] <y1ds> it is the rootdn in the cn=config
[14:53] <ahasenack> ok, then you need an ldapmodify operation on cn=config
[14:53] <ahasenack> something like
[14:53] <ahasenack> sudo ldapmodify -Y EXTERNAL -H ldapi:///
[14:53] <ahasenack> then feed it
[14:53] <ahasenack> dn: cn=config
[14:53] <ahasenack> changetype: modify
[14:53] <ahasenack> replace: olcRootPW
[14:53] <ahasenack> olcRootPW: <newhash>
[14:54] <ahasenack> I think the dn in this example is wrong, locate the right one by dumping with slapcat -n 0
[14:56] <ahasenack> y1ds: https://pastebin.ubuntu.com/p/SBZMKGP7Qq/ that set the new password to the plaintext value of "secreT" (no hash: don't do that)
[14:57] <y1ds> https://pastebin.ubuntu.com/p/zdyQvR2xfk/
[14:58] <ahasenack> your rootdn is just cn=config?
[14:58] <y1ds> apperently
[14:59] <y1ds> i must say i dont know much about ldap, and didnt set this up, im just trying to change some things
[15:00] <ahasenack> see if this works: https://pastebin.ubuntu.com/p/xHYJjPnh9P/
[15:00] <ahasenack> note I used slappasswd to generate a hash for "newsecret"
[15:00] <ahasenack> use whatever password you want
[15:01] <ahasenack> ah, use sudo ldapmodify, not just ldapmodify, of course
[15:01] <y1ds> yes
[15:04] <y1ds> insuficient access
[15:06] <ahasenack> check if the acls are about uidnumber=0+gidnumber=0, that's when root connects to the ldapi socket and is what we are doing here when calling ldapmodify with sudo and -H ldapi:///
[15:06] <y1ds> but is my assumption that i need to authenticate as the rootdn from olcDatabase={0}config,cn=config incorrect?
[15:07] <ahasenack> the acls under cn=config
[15:07] <ahasenack> well, depends on the acls
[15:07] <y1ds> uhm how do I check that
[15:07] <ahasenack> look for attributes olcAccess
[15:07] <ahasenack> for example, this is what allows that uidnumber=0+... to manage the db:
[15:07] <ahasenack> olcAccess: {0}to * by dn.exact=gidNumber=0+uidNumber=0,cn=peercred,cn=extern
[15:07] <ahasenack>  al,cn=auth manage by * break
[15:08] <ahasenack> in my example db
[15:08] <y1ds> okay im gonna check
[15:08] <ahasenack> access to * by thatgui manage
[15:08] <ahasenack> "thatguy" being what ldapwhoami returned, in this case
[15:09] <y1ds> hm I see olcAccess: {0}to *  by * none under dn: olcDatabase={0}config,cn=config
[15:15] <ahasenack> it's the last entry that wins
[15:15] <ahasenack> I mean, if there was no match before
[15:16] <ahasenack> so I guess it's the first one that wins
[15:16] <ahasenack> you go from specific to generic
[15:24] <y1ds> hm okay well im calling it a day for now thanks for the tips :)
[15:25] <ahasenack> y1ds: cheers, good luck
[16:41] <ahasenack> kstenerud: remember you have to explicitly ask for sponsorship in the MP before someone will just tag and upload a package for you
[17:01] <kstenerud> ahasenack: So I just write "please sponsor this"?
[17:01] <ahasenack> something like that
[17:01] <ahasenack> I'd also add the git hash, confirming that's what you want sponsored
[17:01] <kstenerud> ok
[18:08] <ahasenack> rbasak: looking at https://bugs.launchpad.net/ubuntu/+source/mysql-5.5/+bug/1745185 that person could be using that other mysql package that has no "initscripts", right?
[18:08] <ahasenack> wouldn't be solving the bug, though
[18:09] <rbasak> ahasenack: https://bugs.launchpad.net/ubuntu/+source/mysql-5.7/+bug/1592669 is related
[18:10] <rbasak> ahasenack: apart from that bug (which we've fixed now, not sure about what release the reporter is using), I think that correct behaviour and any bugs in postinsts attempting to restart masked services is down to debhelper.
[18:11] <rbasak> ahasenack: I'd like to know exactly what Ubuntu release and package version the reporter is reporting against, and exactly how (what command) the reporter is disabling the service. Ie. full steps to reproduce. Until then, Incomplete IMHO.
[18:12] <ahasenack> that was mysql 5.5, command was systemctl disable mysqld
[18:12] <ahasenack> which indicates it was...
[18:13] <ahasenack> trusty
[18:13] <sdeziel> systemd and trusty?
[18:13]  * ahasenack wonders that
[18:13]  * ahasenack does a quick check on bionic fwiw
[18:14] <sdeziel> bionic didn't ship mysql 5.5 AFAIK
[18:14] <rbasak> I think that's why he's wondering :)
[18:15] <ahasenack> I meant I will try a sequence of disable + upgrade on bionic
[18:15] <sdeziel> oh upgrades could explain the oddity
[18:21] <ahasenack> worked just fine in bionic
[18:22] <ahasenack> it got started up again during the pkg upgrade, but was stopped at the end
[18:24] <ahasenack> failed in xenial
[18:24] <ahasenack> which also has mysql-server-5.7
[18:45] <chillage> Good time, i wanna wondering maybe someone have appboxes.co server or maybe some one have server something like appbox?
[18:51] <ahasenack> I don't
[19:01] <chillage> You know, my friend got rented server for him from him admins.. ant we don't know how open port at that specific "appbox" (Ubuntu 18.10 VNC) for znc, there are not default firewall i think there are specific firewall..
[19:02] <chillage> specific firewall
[19:08] <teward> chillage: well if there's no firewall that you can manage on the server, it'd have to be form the appbox control panel
[19:08] <teward> and if there's not a firewall there you can control then you have to contact the admins
[19:15] <chillage> i'll ask him about control panel, he told me that are desktop and terminal window.. contact with admins who give him server it's little bit dificult because they can't know about znc.. se if we don't found control panel then we try write to appboxes.co a
[19:17] <chillage> so thank you for help:) nice halloween night then, until
[19:26] <plm> Hi all
[19:42] <chillage> good time, plm,
[19:44] <chillage> ok, i'm away, good luck
[19:59] <teward> chillage: if they can't know about ZNC, that suggests that it's not permitted on their infrastructure, so you should avoid running it there.
[20:07] <memphisto> hi, how do i share printscreen here
[20:07] <sdeziel> ahasenack: I too don't know what acpid is used for these days
[20:07] <sdeziel> ahasenack: I've be purging this on servers since Xenial IIRC
[20:07] <ahasenack> heh :)
[20:07] <ahasenack> that's good info :)
[20:10] <memphisto> i can't pass the filesystem setup screen , i'm doing manual partition with LVM, /(root)as btrfs , ext4 home...
[20:10] <memphisto> i'd like to share screenshot but don't know how/where to uplaod
[20:11] <ahasenack> rbasak: (for whenever you read this) is linux-meta (src) a known git-ubuntu import failure? It's a main package, but not imported
[20:11] <ahasenack> memphisto: you should upload it somewhere, there are some free image hosting services
[20:12] <ahasenack> or even dropbox and the like
[20:12] <ahasenack> google drive
[20:12] <memphisto> thanks
[20:12] <memphisto> https://imagebin.ca/v/4L1qdANGnY4j
[20:15] <sdeziel> ahasenack: I dig a little and found a commit in my puppet stuff: "remove acpid from Xenial VMs (not needed thanks to systemd)" dated Oct 17 10:15:24 2017
[20:15] <ahasenack> memphisto: is this 18.04 or 18.10?
[20:16] <memphisto> ahasenack: 18.04.1
[20:16] <ahasenack> sdeziel: I tried a trusty vm, but could still power it off externally without acpid, so I wasn't sure what was going on there
[20:17] <ahasenack> memphisto: that looks like a bug, let me give it a try
[20:17] <memphisto> ahasenack: yes, it looks like it
[20:18] <ahasenack> memphisto: have you tried 18.10 also?
[20:18] <sdeziel> ahasenack: interesting, found another note where I concluded that 15.04+ only needed the dbus package to properly handle ACPI signals in KVM
[20:18] <ahasenack> sdeziel: interesting, I saw dbus messages in the console when I pressed the power button
[20:18] <memphisto> ahasenack: no, i haven't...i like using only lts
[20:18] <sdeziel> ahasenack: I may have missed that 14.04 was also OK
[20:18] <ahasenack> memphisto: that's fine, just checking if it was perhaps fixed in the 18.10 installer
[20:20] <ahasenack> memphisto: the whole disk is a pv, right
[20:20] <memphisto> ahasenack: yes
[20:24] <ahasenack> memphisto: yeah, same here
[20:25] <ahasenack> memphisto: let me try 18.10 to compare
[20:25] <memphisto> ahasenack: great, thanks ... for a moment i thought i'm going crazy
[20:27] <ahasenack> I think it might be because /boot (part of / in this case) is not in its own partition
[20:27] <ahasenack> verifying that oo
[20:29] <memphisto> yesss
[20:30] <memphisto> does it really have to be separate?
[20:31] <ahasenack> don't know yet
[20:31] <ahasenack> the moment I add an actual partition for /boot, I can't setup lvm anymore
[20:34] <ahasenack> yeah, it was /boot
[20:34] <memphisto> ok, how did you create /boot and lvm ?
[20:34] <ahasenack> so I think it's bug https://bugs.launchpad.net/subiquity/+bug/1785332
[20:34] <ahasenack> memphisto: I created a big partition to hold the PV, and left it "unformatted"
[20:35] <ahasenack> memphisto: then created the PV in that unformatted partition
[20:35] <ahasenack> after that I was able to add a /boot partition with the remaining space, outside the pv
[20:35] <ahasenack> and add the lvs to the vg as usual
[20:35] <ahasenack> let me share a screenshot
[20:35] <ahasenack> but the order was important, because the moment you create a /boot partition, the lvm option becomes grayed out
[20:36] <memphisto> no need, got it
[20:36] <memphisto> doing it right now
[20:36] <ahasenack> memphisto: http://people.ubuntu.com/~ahasenack/partitioning.png fwiw
[20:42] <UsQUE> Anyone did migration from FreeBSD samba dc to Ubuntu Server Samba dc?
[20:45] <ahasenack> any sort of samba migration sounds hard on its own
[20:47] <UsQUE> nah I think its pretty easy :P
[20:47] <UsQUE> just backing up the correct files and restore them back on the new system
[21:25] <UsQUE> anyone got HyperV server + Samba DC constrained deligation working ?
[21:50] <compdoc> Hyper-V is too slow for my guests
[21:52] <nacc> UsQUE: isn't that  question for either samba or hyperv?
[21:52] <nacc> (/ microsoft)
[23:32] <RoyK> !ask | UsQUE
[23:36] <rbasak> ahasenack: it's blacklisted.