[01:52] <deadevilboy> hi there guys
[01:52] <deadevilboy> I am using ubuntu server on raspberry
[01:52] <deadevilboy> but when I was upgrading kernel
[01:52] <deadevilboy> an error occurred.. now I can't install or remove anything using apt
[01:53] <sarnold> can you pastebin the terminal output of the error?
[01:53] <deadevilboy> yes
[01:54] <deadevilboy> ubuntu@ubuntu:/boot$ sudo apt autoremove
[01:54] <deadevilboy> freed.
[01:54] <deadevilboy> (--remove):
[01:54] <tomreyn> !pastebin
[01:54] <deadevilboy> sorry guys
[01:55] <deadevilboy> I was using webchat freenode
[01:55] <deadevilboy> I forgot u may be using mIRC
[01:56] <deadevilboy> sarnold https://paste.ubuntu.com/p/4nvdKGcsBT/
[01:57] <sarnold> wow, I don't think I've ever seen that one before
[01:58]  * TJ- puts hand up ... I have!
[01:58] <sarnold> EUCLEAN 117 Structure needs cleaning
[01:58] <deadevilboy> I tried to remove old kernels
[01:58] <deadevilboy> but the error is always the same
[01:58] <sarnold> well whattayaknow...
[01:58] <sarnold> TJ-: hopefully in the context of this error? :)
[01:59] <TJ-> If I recall correctly its due to the 'newish' creation of the symlinks in /boot/{vmlinux,initrd.img}{,.old}
[02:00] <sarnold> deadevilboy: is there anything in your dmesg about your storage? somehow I think of arms as always using sdcards and those usually seem to turn to garbage pretty quick
[02:00] <sarnold>     if (!rename("$dest.$rand", $dest)) {
[02:00] <sarnold> that's the syscall that failed
[02:01] <deadevilboy> maybe this:  EXT4-fs error (device mmcblk0p2): ext4_lookup:1702: inode #13: comm linux-update-sy: deleted inode referenced: 2878
[02:01] <deadevilboy> ?
[02:01] <TJ-> deadevilboy: can you show us "pastebinit <( ls -latr /boot/ )"
[02:02] <sarnold> deadevilboy: yeah that looks like it. maybe that just needs a fsck after a reboot to set everything straight; maybe it's failing media.. are there other lines in the logs that look like block storage problems?
[02:02] <TJ-> deadevilboy: actually, add in the inode flag: can you show us "pastebinit <( ls -ilatr /boot/ )"
[02:02] <deadevilboy> https://paste.ubuntu.com/p/FSXKjRgKB6/
[02:03] <sarnold> oww
[02:04] <TJ-> there's the issue, a hanging symlink from vmlinuz and no updated kernel version
[02:04] <deadevilboy> https://paste.ubuntu.com/p/NBMrBZWTwR/
[02:04] <deadevilboy> I guess with -i won't change the result
[02:05] <deadevilboy> yes, I ran sudo apt upgrade
[02:05] <deadevilboy> and was upgrading from version 1017 to 1018
[02:05] <deadevilboy> then it stopped
[02:05] <TJ-> deadevilboy: you'll need to run an fsck -f on that file-system
[02:05] <deadevilboy> I can't remember the error
[02:06] <sarnold> all those ??? bits means there's something pretty wrong
[02:06] <sarnold> I don't know how to get the machine to fsck itself, maybe it'll do that when booting; if not it might be worth pulling to storage and fscking from another machine
[02:07] <deadevilboy> this is a problem when using sd cards I guess
[02:07] <deadevilboy> everytime the power goes off
[02:07] <deadevilboy> something gets f*
[02:07] <TJ-> deadevilboy: what does "mount | grep boot" report? a 'vfat' file system?
[02:07] <sarnold> yeah they seem really brittle
[02:08] <deadevilboy> I will have to try an M.2
[02:08] <deadevilboy> or an ssd
[02:09] <deadevilboy> yep, I can't fsck from here
[02:09] <deadevilboy> I am connected through ssh at the moment
[02:09] <deadevilboy> away from home
[02:09] <deadevilboy> I will try to do it later on then
[02:09] <deadevilboy> and I will give u my feedback
[02:10] <TJ-> deadevilboy: you can "sudo umount /boot" then "sudo fsck -f /dev/mmcblk0p2" then "sudo mount -a"
[02:11] <deadevilboy> umount: /boot: not mounted.
[02:11] <tomreyn> and maybe replace some of the zeroes in the fstab pass_no column, too
[02:12] <TJ-> deadevilboy: ok, so /boot/ isn't a separate FS?
[02:12] <deadevilboy> I guess not
[02:13] <TJ-> deadevilboy: I stick with the Raspbian distro for a reason when it comes to the Pis ... I found the Ubuntu install brittle and half-baked and hard to reason about its layout
[02:14] <deadevilboy> TJ- maybe u are right
[02:14] <deadevilboy> when there is a new update
[02:14] <deadevilboy> something may get broken
[02:14] <deadevilboy> using raspbian, everything is tested for raspberry
[02:15] <TJ-> The whole Ubuntu-core vs regular Ubuntu confused me for some reason, and that is HARD to do :P
[02:16] <TJ-> We're deploying lots of Pis as office automation controllers, with ZigBee etc., and Raspbian is much more stable and predictable
[02:23] <deadevilboy> I will follow TJ- advice
[02:37] <deadevilboy> thks guys
[02:37] <deadevilboy> for everything
[02:38] <sarnold> bye deadevilboy :)
[02:38] <deadevilboy> sarnold ;)  thank u
[07:22] <lordievader> Good morning
[09:57] <ferrus> Hi.  Is anyone online with experience installing Ubuntu to multipath volumes?
[09:57] <ferrus> Was referred here from the main #ubuntu channel
[09:58] <ferrus> I'm trying to install Ubuntu onto a diskless Cisco UCS blade, with a multipath FC SAN boot volume presented to it.  This has worked fine for RHEL/CentOS many times - but the Ubuntu installer doesn't appear to natively detect multipath.
[10:25] <mwhudson> ferrus: which version?
[10:30] <ferrus> 18.04.4.  Veeam (Enterprise backup product) have started supporting XFS with fast clones, and they're only recommending Ubuntu for it - not RHEL/CentOS, because of the release schedules.
[10:31] <ferrus> I've installed Ubuntu a thousand time at home and on work VMs, and RHEL/CentOS a lot at work - but this is my first Ubuntu Bare-Metal install.
[10:34] <mwhudson> ferrus: hm well it ought to work but it's very late for me :/
[10:35] <mwhudson> ferrus: could you try the latest focal daily from http://cdimage.ubuntu.com/ubuntu-server/daily-live/pending/?
[10:35] <mwhudson> ferrus: not necessarily to install but to see if the disks show up at least
[10:35] <mwhudson> ferrus: oh wait, don't think multipath-tools is installed by default in bionic
[10:36] <mwhudson> ferrus: open a shell and "sudo apt install multipath-tools; sudo systemctl restart snap.subiquity.subiquity-service.service" and see if that helps?
[10:37] <ferrus> Sorry - got disconnected.  Back online.
[10:38] <lordievader> Last time I tried the server install (long time back, mind you) the alternate version had much more tools along this line. As in LVM wasn't supported by the regular installed, in the alternate version it was supported.
[10:40] <ferrus> Yeah - I've already dropped down from the desktop installer because of LVM.  I could understand the lack of multipath on a desktop home installer (though not necessarily manual LVM config) - but it's a suprising ommision on an installer sepcifically marked as 'server'.
[10:42] <ferrus> I followed the - brief, instruction here - https://ubuntu.com/server/docs/device-mapper-multipathing-setup, and put 'install disk-detect/multipath/enable=true' in the boot options, but it doesn't appear to make any difference.  There's no /dev/mapper/mpath volumes presented, they still appear as four distinct devices.
[10:43] <ferrus> I've not much experience with editing the boot options though - so I'm not sure of I've got the syntax/placement correct.
[10:44] <lordievader> <ferrus "Yeah - I've already dropped down"> Are you now using the 'live-server' image or the 'server' image? (http://cdimage.ubuntu.com/releases/19.10/release/)
[10:47] <ferrus> Good point - let me check ...
[10:53] <ferrus> Aha - just checking the differences between the two ISO's, and this is in the release notes - "N.B., If you require multipath you will want to continue to use the alternate installer"
[10:53] <ferrus> Good find - thanks!
[10:58] <lordievader> Hey, there you go 😉
[13:47] <unrecovered> heya! i have an odd problem. i'm trying to set up an ubuntu 16.04 on a virtual machine. host is fedora, networking is done via bridge. so my problem is that ubuntu cannot get an IP. i have several CentOS7 vms on this host and they work fine, so i kind of puzzled what the issue could be
[13:47] <unrecovered> VM is qemu
[13:56] <cpaelzer> unrecovered: if you log into the guest console and explicitly refresh dhcp like "sudo dhclient -r" what does it get/report?
[13:57] <cpaelzer> also if you happen to have control about your dhcp server it is often useful to trace the log there - e.g. check if the dhcp request is coming in and if it is answered
[13:58] <unrecovered> cpaelzer i tried manually start dhclient and result is the same - it cannot get address
[13:58] <unrecovered> though if i manually assign address it seem to work somehow
[13:59] <cpaelzer> unrecovered: so do you have control (or know the admin who does) of the dhcp server and check if the request arrives there?
[13:59] <unrecovered> i have control over dhcp but i highly doubt any request ever reached it... never hurt to check though
[14:02] <unrecovered> ...no, nothing
[14:03] <cpaelzer> ok so nothing reaches dhcp, does the guest say it has a link (throw 'ip link' and 'ip addr' to a pastebin maybe)?
[14:04] <unrecovered> can i throw a screenshot? no network on it (:
[14:04] <cpaelzer> sure
[14:04] <cpaelzer> I mean we just compar the common things one by oen -  a screenshot will do as well
[14:05] <unrecovered> https://ibb.co/54fpkB8 this one is from ifup
[14:08] <unrecovered> cpaelzer https://ibb.co/qMSJxzm ip addr
[14:09] <unrecovered> cpaelzer https://ibb.co/RpMvsYB ip link
[14:11] <unrecovered> i'm not fully familiar with ubuntu networking - i'm more of a rhel guy :D  can something block those requests? or can there be some default routes or something?
[14:11] <cpaelzer> not in a default install
[14:11] <cpaelzer> and nothing on your screenshots looks suspicious yet
[14:11] <unrecovered> problem seem really odd, since i can manually add an address and it kind of works
[14:12] <cpaelzer> and I assume on the host/bridge side it is bridged "the same" as the other guests?
[14:12] <unrecovered> of course
[14:12] <cpaelzer> lbivirt can add iptables rules to things, but I don't see why it would do so different for this particular guest
[14:14] <unrecovered> no iptables rules on a guest...
[14:14] <cpaelzer> the ones by libvirt woudl be on the host
[14:14] <cpaelzer> I'd not expect any on the guest
[14:14] <unrecovered> hum
[14:14] <unrecovered> let's see
[14:15] <unrecovered> damn that firewalld makes iptables unreadable :D
[14:16] <cpaelzer> I'm on the run soon, but the next step in debugging would be to assign it to libvirts default network instead of the bridge
[14:16] <cpaelzer> in case it works there you can ignore the guest portion and focus on the host (at least)
[14:17] <unrecovered> ah
[14:30] <unrecovered> cpaelzer damn, dns also won't work... =\
[14:30] <unrecovered> something's fishy
[15:45] <coreycb> jamespage: thoughts on adding libvirt-qemu to the nova group?
[15:47] <coreycb> jamespage: that would allow libvirt to access /var/lib/nova/instances and we can keep the umask tight
[15:48] <coreycb> might be brittle with other hypervisors though
[15:51] <jamespage> coreycb: in principle I think that's fine
[15:51] <jamespage> would we add that from nova-compute-libvirt?
[15:52] <jamespage> might be tricky as I think the libvirt daemon runs as a specific group
[15:52] <jamespage> but probably do-able still
[15:53] <coreycb> jamespage: I think we could do it in nova-compute-libvirt.postinst
[15:54] <jamespage> +1
[15:54] <teward> rbasak: powersj: fail2ban doesn't have any default nginx logging enabled, but it has some patterns in there that should "just work" given that <HOST> still works.  I have not tested IPv6 *but* the system seems like it was able to convert ::ffff:1.2.3.4 into 1.2.3.4 for the banlist and properly added it to `iptables` directly (not `ip6tables`).  Forgot that I don't have v6 enabled on my containers in testing so I'm going ...
[15:54] <teward> ... to set up a test subnet with IPv6 on my system here to actually *test* if it works right for v6
[15:54] <teward> rbasak: powersj: this is *with* sdeziel's suggestion in place
[15:55] <teward> so specifically wrt f2b this 'should work' sdeziel's suggestion
[15:55] <teward> not making any suggestions to change 20.04 just giving you research results so far
[15:55] <powersj> teward, nice - thanks for the update
[15:55] <teward> PITA to get configured right
[15:55] <teward> but otherwise... :P
[15:55] <rbasak> Thanks!
[15:56] <sdeziel> yeah, thanks, might worth adding to the LP
[15:56] <teward> sdeziel: currently digging into my email gateway so can't check atm
[15:56] <teward> or add to it
[15:56] <teward> (E:EmailDead)
[15:57] <sdeziel> I just checked and logcheck doesn't have rules for nginx so it wouldn't cause any spam due to those embedded v4 in v6 ;)
[20:22] <shubjero> Anyone know how to roll back a package update in Ubuntu 18.04? For example I upgraded to neutron-common 14.0.4-0ubuntu1~cloud1 from 14.0.2-0ubuntu1~cloud0 and would like to go back. This was just a testing system. I tried "apt-get install neutron-plugin-ml2=14.0.2-0ubuntu1~cloud0" but that doesnt seem to work.
[20:23] <shubjero> Maybe you know coreycb ? :)
[20:23] <shubjero> I came across this bug which has coreycb's name all over it: https://bugs.launchpad.net/ubuntu/+source/neutron/+bug/1859649
[20:24] <shubjero> The jist of it is the control plane must not be running an older minor version of neutron.
[20:27] <coreycb> shubjero: that should work, what you tried
[20:28] <shubjero> I get "E: Version '14.0.2-0ubuntu1~cloud0' for 'neutron-plugin-ml2' was not found"
[20:28] <coreycb> shubjero: for the bug I think server-side just needs to get upgraded before client-side
[20:28] <shubjero> coreycb: yeah I was just trying to proactively test some updates im applying to the whole cluster tomorrow.. i like to test on a vacant/test compute node beforehand..
[20:34] <coreycb> shubjero: ah right because 14.0.2 isn't available anymore once you update your cache
[20:35] <shubjero> yeah, shit out of luck
[20:35] <shubjero> repo's dont keep that version at all?
[20:37] <shubjero> I should rethink my package management strategy, haha
[20:45] <coreycb> shubjero: not ideal at all but if you really needed to you could get debs from the staging PPA https://launchpad.net/~ubuntu-cloud-archive/+archive/ubuntu/stein-staging/+packages?field.name_filter=neutron&field.status_filter=superseded&field.series_filter=
[20:52] <shubjero> that could come in handy here, thanks coreycb
[21:58] <sarnold> shubjero: if the archives have removed the version you were using, check /var/cache/apt/archives -- you may still have the old version laying around
[23:20] <theborger> hello guys need help. trying to add a second nic card to 18.04 and setup bonding. but i cant get the 2nd nic card to show up in ifconfig. all the googling says to issue auto "InterfaceName" that does not work
[23:21] <sarnold> that sounds a bit like an /etc/network/interfaces directive, but 18.04 LTS uses netplan to render systemd-networkd configuration by default, no? did you disable that?
[23:22] <theborger> i have not done anything. like i said all the sites say to issue auto "InterfaceName"
[23:24] <genii> !netplan
[23:26] <oerheks> https://netplan.io/examples#configuring-interface-bonding ~~~ https://www.snel.com/support/how-to-set-up-lacp-bonding-on-ubuntu-18-04-with-netplan/
[23:28] <theborger> oerheks: that first link is what i am using for the bonding. but when i issues the clear && echo command only one interface shows up
[23:28] <theborger> enp0s25
[23:28] <theborger> lotec25@hellgate_dell:~$
[23:29] <sarnold> you may need to use ip l to see the other nic
[23:29] <theborger> that is what i get with the command
[23:30] <sarnold> what does ip l say?
[23:30] <theborger> so i just add the 2 interfaces enp0s25 and enp2s0 ?
[23:30] <oerheks> does lspci give a clue?
[23:31] <sarnold> heh, that guide assumes there are exactly two NICs on the machine. not a great guide.
[23:31] <theborger> https://pastebin.com/GgjkUWhg
[23:33] <theborger> so i just set it up with those 2 interfaces? right?
[23:33] <oerheks> the interface is just down.
[23:33] <theborger> oerheks: how can i bring it up?
[23:33] <theborger> that is what i was trying to do
[23:34] <theborger> it all has to be through netplan?
[23:35] <sarnold> if you choose to write systemd-networkd configs yourself, or use ifupdown, you'll have to make sure netplan doesn't overwrite your configs; and there's a handful of packages that don't work with ifupdown
[23:35] <theborger> ok ill follow this guide see if i can get it working
[23:41] <theborger> one ?
[23:41] <theborger> do i need to set the card up first before doing the bonding?
[23:43] <sarnold> I don't have enough network equipment to do bonding myself, but I think you don't set up the individual nics, just set up the bond
[23:43] <theborger> ok