[02:33] <jayjo> I'm finally having some success with installing pfsense - I can see the interfaces now, which is very exciting. I have this ASCII art error that is mentioned in this question, where everything displays as q's. Is that a common mistake? https://serverfault.com/questions/953283
[03:02] <jayjo> could that have to do with what console I chose when installing pfsense? There were four options, ansi, vt100, xterm, and cons25w
[03:02] <jayjo> I did try most of them, the only one I had success with was vt100 although it was not right. The others were even more wrong. Do I need to connect to the VM a certain way in order to use the console correctly this way?
[03:08] <jayjo> My attempt: `virt-install --virt-type kvm --name pfsense --vcpus 2 --hvm --disk format=raw,readonly=on,path=/home/jayjo/pfSense-CE-memstick-serial-2.4.5-RELEASE-p1-amd64.img --disk /home/jayjo/kvm/images/pfsense.qcow2,bus=virtio,size=10,format=qcow2 --network bridge=br_wan --network bridge=br_lan --graphics none --noautoconsole --os-type linux --os-variant freebsd11.3 --boot hd --autostart`
[04:54] <cpaelzer> rbasak: let me re-check the bug ref
[04:54] <cpaelzer> arr, yes it referred tot he former SRU
[04:55] <cpaelzer> I ahve a correct bug but youare right the ref was wrong
[05:00] <cpaelzer> rbasak: https://bugs.launchpad.net/ubuntu/+source/open-vm-tools/+bug/1892266/comments/7
[05:00] <ubottu> Launchpad bug 1892266 in open-vm-tools (Ubuntu Focal) "open-vm-tools 11.1.5 has been released" [Medium,In progress]
[06:05] <lordievader> Good morning
=== JanC_ is now known as JanC
[11:05] <zetheroo> I am seeing this https://pastebin.com/E3wWjWdL logrotate error from a couple systems and am trying to get the mysql-server logrotate file to look at /root/.my.cnf for credentials but can't seem to figure it out
[11:07] <zetheroo> the mysql-server file looks like this https://pastebin.com/yx5cH0dk
[11:08] <zetheroo> and I see it's referring to /etc/mysql/debian.cnf for creds, but there's no password in there and it says "# Automatically generated for Debian scripts. DO NOT TOUCH!" in the header of the file.
=== franksmcb_ is now known as franksmcb
[12:50] <zetheroo> regarding my question about logrotate and mysql-server ... did I miss anything? (had an Internet drop-out :/ )
[12:51] <TJ-> zetheroo: I don't see any chatter since your client disconnected
[12:58] <zetheroo> TJ-: ok thanks
[14:37] <jayjo> I'm trying to run freebsd (pfsense) in a KVM+qemu VM but I keep getting a kernel panic after a couple of minutes of operation. "Fatal trap 9: general protection fault while in kernel mode" - could this error be because of my virtualization type selected? This is what I used: https://bpa.st/UBNQ , could it be the --virt-type or --hvm paramaters?
[14:40] <jayjo> my very simple netplan for completeness: https://bpa.st/PZTQ - otherwise a fresh ubuntu server installation
[14:57] <teward> jayjo: I think you need the standard ISO format not the memstick instance
[14:58] <teward> jayjo: I've had no problems installing pfSense in QEMU-KVM via virt-manager when using their full ISO and not their memstick image
[15:00] <teward> for the Sense installer
[15:01] <teward> jayjo: from the pfSense download page:  "The CD Image (ISO) Installer is used to create a CD/DVD version used to install on virtual machines or systems with a CD/DVD drive."
[15:01] <teward> so you're using a VM, use the CD/DVD ISO first
[15:05] <teward> jayjo: that said, pfSense support is in #pfsense ;)
[15:05] <jayjo> That was my initial attempt, but I was trying to install using the serial console, so I had switched to the memstick.
[15:06] <teward> the memstick needs to be written to a disk first - it's not designed for direct-boot use
[15:06] <teward> if you're working with VMs, you'd need the ISO, but again you need #pfsense for support with pfSense ISOs, etc.
[15:06] <teward> that's a bit beyond the scope of here ;)
[15:06] <jayjo> Thanks for your help - I wasn't sure if it was how I was setting upt the VM. I'm learning a lot but this is rather new still. I'll go to #pfsense
[15:08] <teward> jayjo: they may have the better solution to boot to the serial console, but I don't know if you're going to be able to in qemu-kvm.  Might need to run it with VGA mode on a VNC server mode and then VNC into it remotely, but not sure that'll solve your problems either
[15:08] <teward> i did just boot the latest ISO in qemu-kvm though (GUI, but...)
[15:13] <jayjo> I have seen after doing some research that most users are using virt-manager on their workstation and connecting to their server to provision pfsense, then it is still using VGA. I went for the serial method first because this is a small headless machine, but I'm going to try the ISO with VNC now
[15:39] <teward> jayjo: just going to point out I was able to use the memstick + serial to work with it, your command was missing some bits to complete though  (virsh is how I connected to the console) as for kernel panics I can't help you there.
[15:42] <teward> ooop nope confirmed your iissue jayjo but that's not an Ubuntu issue that's a Sense issue
[15:42] <teward> might be just the memstick installer
[15:48] <teward> oop no i lied
[15:48] <teward> it booted then hardcrashed
[15:48] <teward> i think the memstick installer is at fault here
[15:49] <teward> yeah 100% the memstick installer
[16:10] <teward> rbasak: can you get someone on the server team (or yourself) to help with testing to reproduce an issue?
[16:10] <teward> with nginx
[16:11] <teward> (focal 20.04)
[16:16] <sdeziel> teward: I'm not from the server team (nor Canonical) but I'd be happy to help if I can
[16:16] <teward> sdeziel: https://bugs.launchpad.net/ubuntu/+source/perl/+bug/1897561 i need replicated with stock NGINX - user in #ubuntu was reporting they get a hard-fail with nginx from the repos straight https://paste.ubuntu.com/p/xz5TCj428K/ but it's an underlying perl issue
[16:16] <ubottu> Launchpad bug 1897561 in perl (Ubuntu) "libperl.so.5.30.0 causes nginx to segfault" [High,Confirmed]
[16:17] <teward> if we can confirm it's a problem, then we may be able to kick someone into actually handling this more rapidly with a fix
[16:17] <teward> becuase if all new nginx installs fail 20.04 with Perl like this, and we didn't know this before, then we're a little bit SOL here
[16:18] <teward> sdeziel: the core problem here is perl / pcre is used for the regex matching on things that's core for location matching
[16:18] <teward> so if that doesn't work nginx won't work
[16:18] <sdeziel> teward: the steps to reproduce are missing from the LP
[16:18] <teward> sdeziel: this i know supposedly it happens completely
[16:19] <teward> sdeziel: ah, small caveat, Mibix who reported this in #ubuntu and lotuspsychje pinged me about upgraded in place 18.04 -> 20.04
[16:19] <teward> i wonder if THAT is causing some perl library problems
[16:21] <rbasak> I don't know where to start with that, sorry.
[16:21] <rbasak> Need steps to reproduce in the bug, as sdeziel says
[16:21] <rbasak> Somebody actually affected should be able to provide those
[16:21] <rbasak> And we can't make progress without that
[16:21] <sdeziel> teward: in a fresh focal container, nginx-core doesn't even pull in libperl5.30
[16:21] <rbasak> (or at least we can't be expected to make progress)
[16:21] <teward> to be fair I think they have leftovers from an old install
[16:21] <sdeziel> teward: ldd /usr/sbin/nginx shows no perl
[16:21] <rbasak> I could guess and try and investigate, but I'm not inclined to work on it that inefficiently!
[16:22] <teward> sdeziel: yeah i'm not sure what they've got that's causing that
[16:22] <teward> except they did an 18.04 -> 20.04 install
[16:22] <teward> s/install/upgrade in place/
[16:22] <lotuspsychje> we need an apport collect on that bug too
[16:23] <sdeziel> maybe they have libnginx-mod-http-perl from nginx-extra?
[16:23] <teward> rbasak: does libpcre rely on libperl?
[16:23] <teward> sdeziel: ooh good point
[16:23] <teward> *checks that*
[16:24] <rbasak> I would comment on the bug with a template blurb asking for steps to reproduce, version information, etc, and mark Incomplete and not worry about it until more information arrives
[16:24] <teward> nope, E:NOREPRO
[16:24] <rbasak> But I won't interfere with whatever teward wants to do
[16:24] <teward> rbasak: well I can't repro it on 20.04, but i am mainly asking if someone can repro it stock
[16:24] <teward> which I can't
[16:25] <teward> 18.04 -> 20.04 could be a factor for a failure case though
[16:25] <teward> in which case I need more disk space :P
[16:25] <rbasak> I'm not interested in trying to reproduce it until the reporters tell me how they expect me to reproduce it. Sorry!
[16:25] <teward> yep
[16:25] <rbasak> If more people were marked as affected I'd try harder
[16:27] <teward> :P
[16:27] <teward> rbasak: well, lotuspsychje poked me so :p
[16:27] <sdeziel> I think you've done your due diligence ;)
[16:27] <teward> yep
=== benpro8 is now known as benpro
[16:27] <teward> *throws a banana peel at lotuspsychje*
[16:27] <lotuspsychje> lol
[16:28] <lotuspsychje> i just forward the users to the right direction
[16:31] <teward> lotuspsychje: aka "not me" when they don't provide steps to reproduce
[16:33] <lotuspsychje> yeah, there's a lot of bugs needing pre-fixes and thats what we do with our little team in #ubuntu-bugs-announce
[16:33] <lotuspsychje> pushing them with more info/steps/apport-collects towards the devs
[16:33] <teward> yep
[16:34] <teward> the only thing I can't rule out is libpcre is linked to libperl but that's beyond my knowledgeset too
[16:34] <teward> if it is then it's possible libperl segfault is something before libpcre
[17:03] <lotuspsychje> teward: is this useful? https://trac.nginx.org/nginx/ticket/1831
[17:03] <teward> *waits for slow Trac to load*
[17:04] <teward> lotuspsychje: yeah that's related, see http://mailman.nginx.org/pipermail/nginx-ru/2019-June/062292.html too (Russian, you'll have to translate to English, which my system does by default)
[17:05] <teward> lotuspsychje: it's missing full config output for us to debug though (`nginx -T`)
[17:06] <lotuspsychje> not handy, but at least we know its an existing bug right
[17:08] <teward> yep
[17:08] <teward> https://trac.nginx.org/nginx/ticket/2044 was the recent Ubuntu bug filed in Trac about it, pointed at 1831 which pointed at Perl
[17:09] <teward> but without minimum reproducible examples (not provided ANYWHERE) it's going to be hard to trace
[17:09] <teward> it also looks like it's specific to `reload` calls not restart calls
[17:09] <teward> which is... odd
[17:10] <teward> hmm... still ERR:NOREPRO
[17:10] <teward> btu the ones i'm seeing are 18.04 complaints
[17:10] <sdeziel> teward: got a reproducer
[17:10] <teward> sdeziel: share minimum examples please
[17:10] <teward> unless it's on a bug i'm not on
[17:11] <sdeziel> teward: I'm updating the bug ATM
[17:11] <teward> sdeziel: the Perl bug?
[17:11] <sdeziel> https://bugs.launchpad.net/ubuntu/+source/perl/+bug/1897561
[17:11] <ubottu> Launchpad bug 1897561 in perl (Ubuntu) "libperl.so.5.30.0 causes nginx to segfault" [Undecided,Incomplete]
[17:13] <teward> yeah i'm subbed to the bug :P
[17:15] <sdeziel> teward: done
[17:15] <sdeziel> on the second reload, it sigsev
[17:16] <lotuspsychje> nice work sdeziel
[17:17] <sdeziel> it was a collaborative effort ;)
[17:17] <lotuspsychje> : )
[17:26] <teward> sdeziel: wait, so it only happens on multiple reloads? o.O
[17:26] <sdeziel> teward: that's my experience so far and also matches what the reporters seem to imply
[17:27] <sdeziel> FYI, I haven't linked to the upstream bug found by lotuspsychje but I think it's indeed the same problem
[17:28] <teward> 4th reload triggered on my side
[17:28] <teward> sdeziel: what's interesting is it does NOT happen on a 'restart' of nginx
[17:28] <teward> only on the active live reloads
[17:28] <teward> which... is interesting
[17:30] <teward> at least until you try and restart it too many times...
[17:30] <teward> Oct 13 17:29:32 focal-tester systemd[1]: nginx.service: Start request repeated too quickly.
[17:30] <lotuspsychje> sdeziel teward the bug filer did his apport-collect
[22:20] <jayjo> maybe I should just let sleeping dogs lie... but, do you suspect writing or converting the memstick data to a qcow2 https://askubuntu.com/a/853273 would solve the memstick needing to write to disk problem? I've done a little more reading on the differences between ISO and the flash data with the link shared yesterday, https://wiki.osdev.org/El-Torito
[22:21] <jayjo> even if that proves fruitless, I'm still working on the VNC virsh connection. But first I'm just trying to get access set on my workstation with WOL. Does the default ubuntu vnc server have preferred clients? I have a macbook to test with on my LAN
[22:21] <jayjo> teward: ^ :)