[01:23] <jmadero> hi all, I'm trying to sync up various systems with three users - I'm having a nightmare of a time getting ownership to work right
[01:28] <pragmaticenigma> jmadero, Synology does appear to have the ability to work as an rsync server directly. Perhaps that might be an avenue to persue?
[01:28] <pragmaticenigma> jmadero, I found this article that sounds really close to what you're attempting to do: https://gnax.io/backup-to-your-synology-nas-with-rsync/
[01:29] <jmadero> pragmaticenigma: hm, not a bad idea. I've never heard of an rsync server before
[01:29] <jmadero> I'll read that real quick
[01:33] <pragmaticenigma> jmadero, I've done it with a few of my machines. Rsync has a server service/daemon option. The purpose is so that when you do remote sync'ing, the rsync application only has to do the analysis of the files on it's side, it then uploads that information to the rsync service, which keeps it's own analysis. Then the rsync server sends back what files need to be syncronized. It's designed to reduce the amount of data required
[01:33] <pragmaticenigma> back and forth to do the file signature checking between the two systems.
[01:36] <jmadero> hm - not convinced this will work still but investigating
[01:39] <jmadero> well, this is surprising, running rsync from synology and using ssh is running incredibly fast
[01:41] <pragmaticenigma> not surprising
[01:41] <jmadero> way faster than my new i7
[01:43] <jmadero> new interesting result, running rsync from Synology with ssh to client results in correct user, incorrect group though
[01:45] <pragmaticenigma> I personally could live with the wrong group, so long as I can access my files
[01:46] <jmadero> issue would be that when it loses group the other users can also access it
[01:47] <jmadero> this is all more an intellectual experiment than anything, I honestly don't care if my wife has access to my home folder :-b but, not being able to resolve this is going to make me lose my mind
[01:48] <pragmaticenigma> that's the only concern i have is if you're litterally syncing the entire home folder as one unit... It would be preferred that you only sync what is within the home folder... that way the root of the home user directory retains it's group value and permissions
[01:49] <jmadero> pragmaticenigma: that's true, then there would be just minor hiccups (for instance when it modifies my ssh keys and makes them unusable with bad permissions)
[01:50] <jmadero> I'm doing a roundtrip sync right now using Synology as the rsync "runner", seeing what happens with my permissions
[01:51] <pragmaticenigma> which is why I personally wouldn't venture down this path. Sure it's nice to have everything sync'd, same backgrounds on all machines etc. But I have always found it more of a pain in the long run
[01:51] <pragmaticenigma> tweaks that you use for one machine, may not work on another, and as you go further down the hole, other things will catch
[01:52] <pragmaticenigma> If it were just a matter of sync'ing music, videos and documents, I would only set those up to be sync'd
[01:52] <pragmaticenigma> settings and config files don't always migrate well from machine to machine
[01:52] <jmadero> yeah, I'm finding that to be true haha but, going to fight giving up a bit longer
[01:53] <jmadero> maybe I'll come up with some stupid hack that works - including the almighty "setPermissions.sh" script that runs after every rsync and just sets everything right haha
[01:53] <pragmaticenigma> and I'm also a big fan of not sync'ing my ssh keys, as if a machine were to become compromised or lost. I only have to remove the key for that computer. I don't have to go to all my other machines and reset them up
[01:53] <jmadero> yeah I actually may exclude the .ssh directory
[01:57] <pragmaticenigma> I wish you luck there jmadero ... I've gotta sign off for now.take care
[01:58] <jmadero> pragmaticenigma: thanks for helping, you've given me lots of ideas to try
[04:56] <Wally> Ubuntu is crashing on first login when I set it up with AD. Not sure why. Replicated it on another machine, all the steps were completed. It works fine after first login
[07:20] <lordievader> Good morning
[09:00] <rbasak> Teikoman_fi: o/ I wrote uvtool
[09:00] <rbasak> There's a PPA if you want --network-config in older releases
[09:01] <rbasak> https://launchpad.net/~uvtool-dev/+archive/ubuntu/master
[09:01] <rbasak> However that's unofficial of course
[09:01] <rbasak> And uvtool is now ported to Python 3, so it's not going to work with old releases that don't have the libvirt Python 3 bindings
[09:02] <Teikoman_fi> rbasak: oh hey, yeah at the time I went bad route and applied adams patch directly but I did however test that the devel side does work too.(I even cloned devel locally yesterday and started trying to figure out what it does command wise)
[09:02] <rbasak> Teikoman_fi: not sure what you mean by porting to virt-manager - if you use uvtool, virt-manager should see the created VMs and you can continue there
[09:03] <rbasak> Do you mean a GUI version?
[09:03] <Teikoman_fi> virt-manager is deprecated as far I know and RHEL is replacing it with Cockpit
[09:03] <Teikoman_fi> probably the gui version
[09:04] <rbasak> Ah I didn't know that
[09:04] <Teikoman_fi> I am glad it was ported to Py3(makes easier to read from what I have deciphered so far) and will see how badly everything breaks if I try update our main dedi to that one :S
[09:05] <lordievader> Is it deprecated? Last version is from july last year.
[09:06] <Teikoman_fi> lordievade: https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/8/html/8.1_release_notes/deprecated_functionality sect:  7.11 mentions it being deprecated
[09:09] <Teikoman_fi> Cockpit is nice tho it lacks features specially on VM side which is one of the reasons I wanted to see what Uvtool is doing command wise
[09:10] <rbasak> Teikoman_fi: uvtool glues three things: simplestreams to maintain a local store of cloud images, cloud-init for getting the cloud images to do the right thing on boot, and libvirt to set it all up.
[09:10] <rbasak> If you have any specific questions I'd be happy to go into more detail
[09:10] <rbasak> Flaky connection here though
[09:11] <lordievader> Not sure if there is a difference, but is this not RHEL deprecating virt-manager in favour of cockpit, rather than virt-manager being deprecated?
[09:11] <rbasak> virt-manager's VM setup process always seemed very installer-oriented to me, as if cloud images aren't a thing
[09:11] <rbasak> Perfect for installer testing, but not so much for the "quick VM please" use case
[09:12] <rbasak> Though that can't be hard to add
[09:13] <Teikoman_fi> yeah and lately I have found myself loving the cloud images more where uvtool kicks in like a champ and well since the network-config functionality was added it pretty much fixed the main issue I had with the dedi provider as they do have some weird network things ongoing
[09:14] <rbasak> Thank you for the compliment :)
[09:15] <Teikoman_fi> I only have 1 odd issue with uvtool but it's something that arp -a + virsh domiflist fixes quickly
[09:16] <Teikoman_fi> if I create a VM that uses additional ipv4 from network-config, for some reason uvt-kvm ip <vm> is unable to get the IP of that box therefor I need to use some superglue awk + grep magic to pull mac from domiflist and try match it with the arp -a response :)
[09:17] <rbasak> Yeah uvtool's IP address determination is heuristic only and doesn't understand any configuration customisation such as adding another IP
[09:20] <rbasak> All we can do is add additional heuristics on a case by case basis
[09:21] <rbasak> Bug 1825263 is for your particular case I think
[09:22] <Teikoman_fi> rbasak: Currently this pasta is what I use(currently broken but more or less for reference) https://paste.ubuntu.com/p/Yr74KRrz3C/ that would list me the VMs in general with their name, mac, ipv4, status and bridge
[09:23] <rbasak> Thanks! Maybe paste that into the bug for others?
[09:24] <Teikoman_fi> idea would be to try replicate what virsh does with `virsh net-dhcp-leases <network>` but with multiple VMs without  having to specify the network or machine specifically.
[09:24] <Teikoman_fi> sure, just need to try fix it as said it's not fully working(only returns 1 VM from 2 for whatever reason)
[09:35] <Teikoman_fi> Now I am just wondering that how safe is it to try get the master ppa as 18.04 shows latest possible is 140 and how badly will the existing machines break after the update since the 140 was patched with adams blog patch file(as it was required for Hetzner to even work)
[10:35] <Teikoman_fi> rbasak: so tried fixing it abit and now it's pretty much this(with the output after #): https://paste.ubuntu.com/p/2n7w9qfc8D/ sadly I am not sure if it's really useful for the Bug 1825263 at all and ofc this does not cover if you have multiple NICs(not tested) plus this is with direct network-config IP mapped VMs to public ipv4 without private network routes.
[11:15] <ackk> hi, I have a netplan question, is this the right place to ask?
[11:40] <lordievader> ackk: If it isn't we'll let you know ;)
[12:22] <ackk> lordievader, :)
[12:23] <ackk> so, I have a container with a single eth0, I added netplan config to create a br0 bridge with eth0 in it (setting the bridge mac to the one of the eth). now if I revert the original config with just eth0, netplan doesn't seem to do anything
[12:24] <oerheks> are you sure it is eth0 ?
[12:25] <oerheks> ifconfig would show the names of interfaces
[12:25] <ackk> oerheks, wdym?
[12:25] <lordievader> OerHeks: ip link show these days ;)
[12:26] <lordievader> ackk: Did you re-apply the netplan config after changing it? (I don't use netplan but vaguely remember that you have to re-apply the config after change)
[12:26] <ackk> https://paste.ubuntu.com/p/R8PzyFPtjP/ is the config I aded
[12:26] <ackk> lordievader, yes. and I actually now noticed that br0 doesn't have an address anymore (but it's still there) but also eth0 didn't get an address
[12:27] <lordievader> ackk: Do you have bridge-utils installed?
[12:27] <ackk> lordievader, https://paste.ubuntu.com/p/TmypGVDnMt/
[12:27] <ackk> lordievader, no
[12:27] <ackk> I thought those were deprecated too?
[12:27] <ackk> ip l del should do?
[12:28] <lordievader> True, they are. Just like the output more than iproute2's version.
[12:28] <ackk> lordievader, +1
[12:28] <ackk> ip makes it very hard to see what's in a bridge
[12:29] <lordievader> ackk: What is the output of bridge link?
[12:29] <ackk> 25: eth0 state UP @if26: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 master br0 state forwarding priority 32 cost 2
[12:29] <ackk> lordievader, ^
[12:30] <lordievader> So eth0 is part of the bridge.
[12:31] <ackk> yes
[12:31] <lordievader> Is br0 set to dhcp?
[12:32] <lordievader> Is there a dhclient running for br0?
[12:32] <ackk> lordievader, well not anymore, I removed it. to be clear, adding the config from the first paste gave me br0 correctly working (with an ipv4 and eth0 in it). what I was trying to do is to revert to just eth0
[12:32] <ackk> lordievader, no
[12:33] <lordievader> Oh, you are trying to remove the bridge interface? And just do dhcp on eth0?
[12:33] <lordievader> In that case read https://wiki.archlinux.org/index.php/Network_bridge#With_iproute2
[12:33] <lordievader> It shows you how to create the bridge, show it, and remove it again.
[12:34] <lordievader> Then define your netplan config how you want it and apply that.
[12:40] <rbasak> Teikoman_fi: yeah that makes sense. Thanks anyway!
[12:45] <Deihmos> etwork installer for the server ?
[12:46] <Deihmos> isn't there a network installer for the seerver
[12:50] <tomreyn> Deihmos: what are your requirements, what are you looking for in a "network installer for the server"?
[12:50] <tomreyn> pxe booting? downloading updates during installation?
[12:50] <Teikoman_fi> Hmmph almost deciphered 1 of the routes from uvtool to pure commands :D Python looks so much more cleaner and idk if bash can even properly generate the config/yaml files so nicely. At least the deciphered way covers my personal use case for ubuntu images. Now to pack it into some sort of sh file to cover those tedious converter + yaml generation steps
[12:51] <Deihmos> so when i install it is already up to date
[12:51] <tomreyn> Deihmos: the installer, or the installed system?
[12:52] <Deihmos> installed system
[12:53] <tomreyn> Deihmos: if i recall correctly both 18.04 LTS server installers take care of this
[12:53] <tomreyn> Deihmos: have you tried and found it not to be so?
[12:53] <Deihmos> no I just did an install and it had 60 updates
[12:54] <Deihmos> installed 18.04 lts server
[12:54] <tomreyn> using which installer?
[12:54] <Deihmos> the standard installer
[12:54] <tomreyn> 18.04.3 live-server?
[12:57] <Deihmos> yes
[12:57] <Deihmos> https://ubuntu.com/download/server
[12:58] <Deihmos> i think i need to use the mini iso to get a network install. problem is the mini is a little different from the server version
[12:58] <pragmaticenigma> Deihmos: Different in what way?
[12:59] <Deihmos> for one it has a splash screen when you boot it. I have to manually disable that stuff
[13:00] <ackk> lordievader, oh ok, so I have to remove the bridge manually. I thought netplan would do it, as the doc mentions at some point that interfaces that get removed from the config are unconfigured
[13:00] <tomreyn> Deihmos: how many systems are you installing?
[13:01] <Deihmos> 3. I will just use the live-server. no big deal
[13:01] <lordievader> Deihmos, tomreyn: There is an update tickbox in the installer right? Given you have an internet connection.
[13:02] <Deihmos> didn't see that tickbox
[13:02] <lordievader> ackk: Well, I suppose it could do it. But like I said, I've never used netplan, so I don't know if it actually will.
[13:03] <Teikoman_fi> well if it's normal server installer it does go through DHCP setup parts etc visually even when configuring the server and if it fails, it asks you if you want to manually set it up(bumped into this like yesterday at least).
[13:03] <tomreyn> lordievader: it's been a while that i used the server installer, i don't want to make any false statements.
[13:04] <lordievader> Hrmm, same here.
[13:07] <tomreyn> last but one time i tried you could live-update the installer itself, i.e. snap refresh subiquity, and benefit from its new features and bug fixes. but last time i tried this was subiquity was longer hosted on snapcraft.io so this was no longer possible.
[13:08] <tomreyn> i think there are nightlies or weeklies now, though, at least for 20.04, maybe 18.04, too.
[13:10] <tomreyn> so things are in flux, as is more or less confirmed by https://discourse.ubuntu.com/t/server-installer-plans-for-20-04-lts/13631
[13:14] <tomreyn> hmm there is http://cdimage.ubuntu.com/ubuntu-server/daily-live/current/ but while the directory name suggests "daily" the amd64 one was last built on january 7.
[13:19] <powersj> tomreyn, those are blocked by tests if you go up a directory you will see there is a pending directory with the ISO from today
[13:19] <powersj> I know paride is looking into the tests now to see what has prevented the latest from getting promoted
[13:19] <paride> -> true
[13:21] <weedmic> can a browser cause a system freeze that ctrl+alt+f3 won't even work?  and if yes, can it also cause a reboot? and if yes, how do I fix it?
[13:21] <tomreyn> powersj: thanks. so the assumption is those can be ok, aynd in contrary to what i was told previously (https://bugs.launchpad.net/ubuntu/+source/subiquity/+bug/1803338/comments/11) live images can generally be used for testing again (at least after they passed tests)?
[13:24] <pragmaticenigma> weedmic: something like that is possible... and when I've encountered it, it's because a web site was using malicious code (e.g. Digital Currency Miner)
[13:25] <tomreyn> weedmic: web browser (i assume) related questions are better placed in #ubuntu, i'd say.
[13:25] <powersj> tomreyn, they are ok, in that they passed a smoke test and are ok for development work; not for production
[13:25] <tomreyn> sure, not for production
[13:26] <weedmic> ok, so far the only reboots on the server were while either FF or Chrome were running.  but I do not think malicious sites were accessed, but can check.  just wanted to know if it is possible.
[13:34] <weedmic> is there a command to see what voltage is actually reaching each cpu/core/thread?  something like htop, but for votage?
[13:37] <paride> tomreyn, powersj, but note that the failed test that prevented the current image promotion in the last days *is the smoke test*, so the latest "current" images is not to be considered ok for anything for the moment
[13:37] <tomreyn> weedmic: maybe the bios canhint on it. other than that there may be proprietary system management hardware providing this infomation, but this is hardware specific.
[13:38] <paride> sorry, I mean the latest "pending" images
[13:38] <tomreyn> paride: thanks for the warning, i've actually settled for testing the latest which passed the test
[13:38] <weedmic> not part of inxi - re read man page
[13:39] <weedmic> q - rebooting to check cmos tests
[13:41] <pragmaticenigma> dang it... they left just as I found the answer... lm-sensors can read CPU temp, voltage, and fan speeds
[13:45] <lordievader> _weedmic: An out-of-memory situation where the machine starts swapping like crazy may also seem like switching tty's is unresponsive. Perhaps your browser was eating all the ram?
[13:46] <pragmaticenigma> lordievader: I think they're offline still
[13:55] <weedmic> this was useful - watch -n 2 sensors
[13:59] <lordievader> _weedmic: Powertop might be able to give some details too as to where the power is going.
[14:03] <pragmaticenigma> weedmic: also, lm-sensors can also view voltages: https://www.cyberciti.biz/faq/how-to-check-cpu-temperature-on-ubuntu-linux/
[14:07] <lordievader> lm-sensors == sensors ;)
[14:08] <pragmaticenigma> yeah... strange that you don't install "sensors" to get it though
[14:10] <lordievader> Well, the package provides more binaries than just sensors.
[14:14] <tomreyn> hmm, got a nice grub shell after installation. :-/
[14:15] <lordievader> Better than "medium not found"
[14:15] <lordievader> ;)
[14:15] <tomreyn> yes, grub installed fine (or maybe it's the one that was already there)
[14:25] <tomreyn> lvm is missing
[14:26] <tomreyn> actually not.
[14:28] <lordievader> Your version of grub is able to boot to lvm directly?
[14:29] <tomreyn> yes it should be,
[14:33] <pragmaticenigma> !coc
[14:34] <pragmaticenigma> sorry folks... wrong tab
[14:50] <distek> Hey all! I'm having an issue with my Ubuntu Server. I'm currently running it in a VM as my Nextcloud instance. It's running quite well for the most part! My issue is that it seems to just stop allowing connections randomly. Any means of accessing it, whether it be via LAN or WAN fails. It appears I can fix it by just resetting the interface from within the VM (ip link set [dev] [down/up]). This is
[14:51] <distek> less than ideal, of course, so I was wondering if anyone has an idea on how to track down the root cause?
[15:07] <tomreyn> distek: is there anything about it on the logs?
[15:08] <tomreyn> which virtualization are you using, how is traffice passed into and out of the VM?
[15:29] <tomreyn> i think i identified the problem with the server installation. i had edited the archive mirror the server installer had chosen, changing http to https, which none or not all of the servers this hostname resolves to actually support.
[15:30] <tomreyn> the installer accepted this change, there was no warning or error about it, the installation was considered successful.
[15:31] <distek> Not that I can see. I'm a bit confused as to what's controlling the network to begin with. NM doesn't appear to be present, so no luck there. Didn't see anything in particular in journalctl or dmesg. The interface appears to be able to reach out, just nothing routes back to it(when it starts this non-sense).
[15:31] <distek> I'm using libvirt/qemu
[15:33] <distek> More info: Interface is using macvtap(private). I'm not concerned with the host being able to directly communicate with the guest. Just the rest of the LAN
[15:34] <tomreyn> distek: inspect interface statistics in the VM and on the host. next time it occurs, ping out of the vm and into it by ip address and watch the traffic ((tcpdump) and compare it to when you do this while the traffic flows fine.
[15:35] <tomreyn> see if you can still use some but not other protocols when it happnes, i.e. icmp vs udp vs tcp.
[15:36] <distek> I'll give it a try! Thanks tomreyn!
[16:21] <ahasenack>   scan: scrub repaired 0B in 0 days 03:43:10 with 0 errors on Mon Jan 20 15:54:18 2020
[16:21] <ahasenack> always good news
[16:36] <tomreyn> bug 1860352
[17:14] <strangezak> Hey guys im a little lost here, so my system is having a CPU lockup when server goes down and i walk over the monitor connected to the server and its flooded with CPU #1 stuck for 22 seconds [(networkd):6019]. So i restart the box and enable verbose logging for systemd-network and im getting this log every minute or so https://pastebin.com/DVXywEme 192.168.0.27 isn't even a valid ip on this network. I checked what is running under the source port
[17:14] <strangezak> and it says dnsmaq, but dnsmasq isn't even installed on this system when i sudo apt-get remove dnsmasq it says dnsmasq is not installed. Any ideas on how i can investigate this further
[17:53] <rafaeldtinoco> strangezak: several other services use dnsmasq for "internal dhcp/dns" stuff (lxd, libvirt).
[17:53] <rafaeldtinoco> it seems you're getting a soft lockup (20 sec) in systemd-networkd
[17:54] <rafaeldtinoco> better path here is to open a bug over systemd in launchpad.net and inform ubuntu version, kernel version, and probably attach more logs to the bug
[17:55] <rafaeldtinoco> https://bugs.launchpad.net/ubuntu/+source/systemd/+filebug
[17:55] <rafaeldtinoco> to help you out ^
[21:31] <Wally> I'm getting the "oh no something's wrong" when a new AD user has logged into my Ubuntu machine. The logs seem to indicate " Check that logind is properly installed and pam_systemd is getting used at login.
[21:32] <Wally> Pretty sure that's all setup properly..
[21:32] <Wally> It works on second login though which is even odder.
[23:03] <Wally> I assume likewise-open is unmaintained
[23:14] <pragmaticenigma> Wally: secong login attept, or different user?
[23:25] <Wally> second logiun attempt
[23:28] <pragmaticenigma> Wally: sounds like something just took longer than expected to perform the authentication