[07:02] <cpaelzer> gz powersj to also officially be part of the group now :-)
[07:23] <cpaelzer> rbasak: once you are aroung it would be great if you could check my ubuntu-release question (if no one else answered) and tell me if/what I overlook
[07:23] <cpaelzer> thanks in advance
[07:23] <cpaelzer> about the MIR related apache2 migration
[07:47] <cpaelzer> rbasak: resolved
[09:49] <azidhaka> Help with 16.04 upgraded from 14.04 :) /var/log/syslog is empty, everything goes to syslog.1. Logrotate config: https://pastebin.com/mAiJVKJY ;  Rsyslog config: https://pastebin.com/3V7BbVX5
[09:56] <rbasak> azidhaka: have you rebooted since the upgrade?
[10:50] <cpaelzer> rbasak: can you pull bug 1735930 into the ongoing merge?
[10:54] <Neo1> Hi! Who know how set up php5.6, Now I have 7.1
[10:54] <Neo1> Used this manual and this doesn't work https://askubuntu.com/questions/109404/how-do-i-install-different-upgrade-or-downgrade-php-version-in-still-supported#109544
[10:55] <rbasak> cpaelzer: added to my list. Thanks!
[10:55] <rbasak> I'm planning on doing the merge first, then a further upload to fix bugs.
[10:56] <cpaelzer> just wanted to know it in good hands
[10:56] <cpaelzer> thanks
[11:28] <jamespage> coreycb: doing oslo.* updates for m2; oslo.log is unhappy under py3.6 atm (applies to current version as well)
[12:19] <jamespage> coreycb: I think we need to revert the exception serialization feature in oslo.serialization; we can drop any patches once upstream either follow suit, or fix the rest of the world to work with it.
[12:46] <coreycb> jamespage: thanks, i'll work on clients today. good call on oslo.serialization.
[14:08] <marlinc> Anyone who has experience with suppressing PAM messages about sudo from a user? I tried to suppress them with "session [success=1 default=ignore] pam_succeed_if.so quiet uid = 0 ruser = zabbix" in /etc/pam.d/sudo but this causes segmentation fault
[14:21] <catphish> i'm attempting to disable dependence on filesystem UUIDs, but i don't seem to be able to stop grub putting them in grub.cfg (specifically in the search line), is there a good solution to this?
[14:26] <catphish> i think the only solution is to hack it out of /usr/share/grub/grub-mkconfig_lib - this feels like a bug
[14:49] <jamespage> coreycb: taking stevedore
[14:49] <coreycb> jamespage: ack
[15:02] <coreycb> jamespage: taking ironic deps and keystoneauth
[15:03] <jamespage> coreycb: not doing everything via bileto - anything pretty atomic can test out in bionic-proposed
[15:03] <jamespage> coreycb: stevedore uploaded
[15:03] <coreycb> jamespage: ok
[15:03] <jamespage> coreycb: ovsdbapp next
[15:40] <powersj> cpaelzer: thanks!
[16:11] <jamespage> coreycb: os-traits done
[16:12] <jamespage> coreycb: python-troveclient next
[16:39] <jamespage> coreycb: done; tempest next
[16:46] <coreycb> jamespage: ack. done up to ironicclient. keystoneauth/middleware and ldappool next.
[16:56] <jlacroix> I'm hoping someone can help me with connecting an external USB hard disk to a KVM/Qemu VM. I've tried installing every package with the word "spice" in the name, on both the host and VM, and I've tried reconnecting, rebooting, etc. For some reason, no matter what I do, the USB drive doesn't show up in the guest VM.
[16:58] <cpaelzer> jlacroix: let me try to find you my last description
[16:59] <cpaelzer> in some bug around that I should have steps to do so ...
[16:59] <cpaelzer> jlacroix: in what Ubuntu release are you?
[16:59] <cpaelzer> (btw no spice needed at all)
[17:01] <cpaelzer> I fixed the remaining apparmor bits in artful, so if you are on 17.10 or 16.04 + UCA-Pike you are good
[17:01] <cpaelzer> otherwise you need to adapt some configs in regard to appamor
[17:01] <cpaelzer> the TL;DR to attach is this:
[17:01] <cpaelzer> 1. make an xml describing your usb device
[17:01] <cpaelzer> 2. run virsh attach-device <guestname> <xml-describing-your-device>
[17:02] <cpaelzer> Bonus: if you need USB2 caps make sure your guest has a usb 2 host controller
[17:03] <cpaelzer> list of usb controllers is at https://libvirt.org/formatdomain.html#elementsControllers if you need more than the default (which strives for max compatibility)
[17:04] <jamespage> coreycb: tempest done - enabled signature verification of the tarballs using the openstack infra signing key
[17:04] <jamespage> neat
[17:04] <cpaelzer> bugs 1552241 and 1686324 have examples how I did test the fixes
[17:04] <jamespage> coreycb: python-os-win next
[17:04] <cpaelzer> jlacroix: I hope that gets it going for you I'm EOD actually so I should leave
[17:05] <coreycb> jamespage: cool will take a look. i think i saw that mentioned in the debian policy upgrade checklist.
[17:05] <boxrick> During my pre-seed postinstall chroot environment. In /lib/modules it has kernel 4.4.0-96 Yet if it I try and use modprobe it looks for 4.4.0-103. Why this difference?
[17:05] <boxrick> And is there any way to stop this happening?
[17:05] <jamespage> coreycb: I've been doing wrap-an-sort -bast to make future merging with Debian a bit easier
[17:07] <coreycb> jamespage: ok
[17:10] <jamespage> coreycb: os-win done
[17:10] <jamespage> coreycb: python-os-brick next
[17:10] <coreycb> jamespage: signing sure is nice
[17:11] <coreycb> jamespage: checking signature that is
[17:14] <boxrick> Any ideas?
[17:23] <ahasenack> boxrick: what's the running kernel?
[17:24] <boxrick> I am using a preseed file, then in the d-i preseed/late_command section running a chroot in the /target
[17:24] <boxrick> So it is the machine I have literally just installed
[17:26] <ahasenack> well, modprobe will consider the running kernel
[17:26] <boxrick> Isn't it just a busybox environment at that point?
[17:27] <ahasenack> and a chroot
[17:27] <ahasenack> loading a module there shouldn't be done I think
[17:28] <boxrick> I need to do some ZFS hacking, so I need something to have loaded in ZFS
[17:28] <ahasenack> but during installation?
[17:28] <boxrick> It was working in a VM, but moving this to real hardware seems to have broken something. Which makes it feel a bit fragile
[17:29] <boxrick> Yea, since I want my /var living on ZFS partition
[17:29] <dpb1> we are doing work in curtin to enable zfs for root, rharper might have more info here. :)
[17:32] <boxrick> I guess I could do something like: d-i base-installer/kernel/override-image string linux-image-4.4.0-96-generic
[17:32] <rharper> generally you'll need to have zfs.ko loaded in your "install" environment; so that looks like 4.4.0-96; so you'll  want to make sure the linux-image-generic-`uname -r` is installed; that'll make sure you have a zfs.ko that can be loaded;  one loaded the zfsutils-linux  is needed for zpool/zfs commands;  lastly, in target/chroot you want zfsutils-linux
[17:33] <rharper> I'm not preseed savvy but maybe someone knows how to do the above requirements in preseed
[17:34] <boxrick> Thats what I was basically doing, but for some reason it was installing 4.4.0-96 then at last minute updating it to 4.4.0-1xx. Which is quite odd
[17:35] <rharper> well, whatever uname -r says, is what's needed, even if you update the install environment to a newer level,
[17:37] <rharper> if you're running a late command, IIUC, and it's in the chroot, then zfs was already loaded and used to create your /var mountpoint ...  so was there an error or was the concern about seeing a different kernel version
[17:38] <boxrick> No, this was a specific error when it was modprobe zfs in the late command, due to kernel inconsistencies.
[17:38] <rharper> it's very likely your install media/environment is 4.4.0-96, but the latest kernel 1xxx is being installed since the installer is going to install the latest version of the package for the target
[17:38] <rharper> oh, you can't moprobe late
[17:38] <boxrick> Well, this is a chroot
[17:38] <boxrick> So its directly in the target
[17:38] <rharper> you have to do it before
[17:38] <rharper> it has to match the installer kernel
[17:38] <rharper> so, if there's an early command
[17:39] <rharper> load zfs early, and then in late command, you can run in-chroot zfs commands as long as the target also has zfsutils-linux installed
[17:40] <boxrick> Early command feels like CentOS. Not sure if thats a thing in pre-seed
[17:40] <boxrick> Can certainly check though
[17:41] <rharper> yeah, I'm not sure on preseed
[17:41] <rharper> looks like there is a di preseed/early_command
[17:42] <boxrick> Plus its some gross busybox environment. You cant really do much
[17:42] <rharper> you only need to modprobe zfs
[17:43] <rharper> if you're waiting until your in your target; though honestly, you should be creating your var before you install the OS , otherwise you'll need to migrate all of the data in var onto your zfs var
[17:43] <boxrick> Thats fine, doing the zfs stuff before install though is a complete nightmare.
[17:43] <rharper> yes
[17:43] <boxrick> Moving some files from a none running os though is fairly easy
[17:43] <boxrick> Create zfs pool, do mountpoint, move files
[17:44] <rharper> if the tools were available, it would save some io
[17:45] <boxrick> Maybe with time this will be easier. I will give this early modprobe a go though, cheers for the hint.
[17:45] <rharper> sure
[17:46] <boxrick> Seems its also getting very confused with disk ordering which is sad
[17:46] <rharper> for zfs or something else?
[17:47] <boxrick> Oh just in general, this is server oddities.
[17:47] <rharper> recommendation is to use /dev/disk/by-id/ links which are stable, where as the sdX values can change
[17:47] <boxrick> It has an onboard SATA card which should be seen first ( sda ) but instead is sdc
[17:47] <boxrick> Can you preseed all that also ?
[17:47] <rharper> I dunno; can you specify the target disk ?
[17:48] <rharper> like, partition /dev/FOO and there, you can replace sda with a link but you need to know it (and it makes it less general)
[17:51] <boxrick> Sadly this is the recipe for countless servers rather than a single one, so it will need to remain sda etc. But thats fine. Will just have to do a specific pre-seed for this group of hardware
[17:52] <rharper> ah
[17:56] <boxrick> Thanks for the help anyway, very much useful info :)
[17:56] <rharper> sure
[18:07] <ScottE> boxrick: just an FYI to make sure you set the "overlay" ZFS option if you have /var on a discrete filesystem, otherwise it will fail to mount at boot since there is some stuff written to /var early on
[18:09] <boxrick> I did find that out earlier the hard way, and ZFS wouldn't mount because of existing files there
[18:48] <ahasenack> rbasak: the other day you said you missed update-maintainer in that mysql merge, did you push that?
[18:48] <rbasak> ahasenack: no I still need to do that.
[18:49] <rbasak> Normally I'd rebase and redo the merge commit, but that would screw up your review.
[18:49] <rbasak> But I suppose I should just add a commit at the end and it'll be fine.
[18:53] <ahasenack> yeah, just add it on top
[19:29] <jlacroix> @cpaelzer sorry I had to step away from my computer and I missed your messages. Ubuntu 16.04 on both the host and the guest
[19:31] <jlacroix> @cpaelzer if it matters I'm using virt-manager to attach the USB device so would I still need to do the xml stuff?
[19:36] <ahasenack> rbasak: still around?