[04:10] <Goop> I have a VPS provider that I would like to use to resell mail and web hosting. The smallest package they provide is $5/mo, and I want to turn around and sell a $5/mo service. What would be the best way to split the virtual hosts/domains between users?
[04:11] <sikun> what would cause a systemctl restart service command to cause a server to hard reboot, IPMI logs aren't showing any entries when they reboot, they reboot almost instantly so I haven't been able to get anything from logs and this is only happening to a handful of the nodes in the cluster.
[04:11] <Goop> I'll be offering Postfix/Dovecot, MySQL, and Apache2.
[04:12] <Goop> sikun, can you give me a context?
[04:14] <sikun> I have a script that I run to restart a service on nodes, it simply runs a remote ssh commmand "ssh mgmt@192.168.1.2 systemctl restart service_name"
[04:15] <Goop> Wait wait wait, that still doesn't tell me the context. What are nodes? virtual machines? SQL nodes?
[04:15] <sikun> as soon as that command finishes, a handful of nodes will instantly power cycle.
[04:15] <sikun> compute
[04:15] <sikun> they all PXE boot, no local storage.
[04:18] <sikun> all are the same model of server, have the same amount of RAM, every single config option in the BIOS are identical, yet.. starting here recently I've been having about a dozen nodes start rebooting on their own, so I'd simply run a script that would "provision" them after they've booted and they'd be good to go but today these nodes started rebooting as soon as the provisioning script finished running the remote SSH commands
[04:18] <sikun> I've dug through their IPMI logs, none show any CPU/RAM/Thermal events
[04:20] <sikun> they reboot whether the command is remotely ran or locally ran
[04:43] <cpaelzer> good morning
[06:10] <lordievader> Good morning
[07:06] <hellyeah> corrupted stack and detected inside scheduler what does that mean?
[07:16] <lordievader> hellyeah: Where does that come from?
[07:21] <hellyeah> when i try to install ubuntu with vmware
[07:35] <lordievader> Is the error from Ubuntu, or from vmware?
[07:45] <hellyeah> lordievader,  i amm not sure about that
[07:45] <hellyeah> but i got this error in ubuntu shall
[07:45] <hellyeah> i mean i booted ubuntu and i saw kernel panic
[07:45] <hellyeah> i guess kernel panic is about ubuntu
[07:46] <lordievader> During the boot of a live-cd?
[07:46] <lordievader> https://askubuntu.com/questions/825841/kernel-panic-cant-boot-into-ubuntu-after-windows-did-some-updates
[07:56] <hellyeah> i boot with wmware
[07:56] <hellyeah> and i got kernel panic
[07:57] <hellyeah> man there is no answer there :D
[07:58] <lordievader> hellyeah: Did the kernel panic occur during the boot of a live image or during the booting of an install?
[07:58] <hellyeah> during the boot of install
[08:00] <lordievader> Does a live-cd boot correctly?
[08:01] <hellyeah> i am not sure
[08:02] <hellyeah> can irt be about windows 10?
[08:04] <hellyeah> it*
[11:20] <vimes> Hello! I run a small ubuntu server hosting some websites for student associations and some small businesses, it runs some django, some php, nginx etc. And it's 16.04, but now 18.04 is here and I've been told "always upgrade", should I? configuring all of this to work takes times, I would re-install everything any how. Is it generally a "good idea" to upgrade to latest LTS or just sitck with 16.04 for it's lifecycle?
[11:22] <hateball> well for starters you can't officially upgrade until 18.04.1
[11:22] <vimes> ooh
[11:22] <hateball> and even then I guess it is up to you. security stuff will keep being patched in 16.04, but you wont get new features for the services you're offering
[11:22] <Nafallo> if it's working and there is nothing new you need, you don't have to upgrade "just because" no.
[11:24] <Nafallo> changing production environments is obviously always a risk you have to decide if it's worth making.
[11:24] <vimes> read a sysadmin book recomending me to always use newest stable versions, but I guess 16.04 will be updated as well, thanks! :D
[11:24] <Nafallo> until April 2021 ;-)
[11:25] <Nafallo> after that you could always upgrade to 18.04 straight to 20.04 :-)
[11:25] <Nafallo> two steps
[11:25] <vimes> It's probably time for me to learn ansible soon any how, then I assume upgrading would go much faster
[11:26] <Nafallo> depends how you do it, but I'd probably wouldn't do an upgrade with ansible.
[11:26] <Nafallo> rather a new deploy/re-provision if so.
[11:26] <vimes> no no, install from scratch, use ansible to install and configure everything
[11:26] <Nafallo> yeah, that.
[11:26] <Nafallo> :-)
[11:46] <rbasak> cpaelzer: could you glance at a shell script for me please? I'd like a second opinion on my review comment for ahasenack's nvdimm packaging.
[11:46] <rbasak> https://git.launchpad.net/~ahasenack/ubuntu/+source/pmdk/tree/debian/tests/manage-pools?h=clean-changelog-for-upload
[11:46] <rbasak> cpaelzer: lines 61 and 68
[11:46] <rbasak> cpaelzer: are the purpose and operation of these obvious to you?
[11:54] <tyx> Hi
[11:55] <tyx>  How to open port# when it in nmap is "filterd" ?
[11:55] <tyx> i'm using ubuntu 16.04 and ufw interface
[11:57] <kiokoman> tyx: example -> sudo ufw allow 2222/tcp
[12:45] <cpaelzer> rbasak: reading
[12:47] <cpaelzer> 61: almost yes
[12:47] <cpaelzer> 61: I see what it is doing, the check is fine and it follow the no-message = good message approach
[12:48] <cpaelzer> but in case the grep goes off and can't find a "consistent" under the reqzested pool_file
[12:48] <cpaelzer> then IMHO it will just break due to the set -e
[12:48] <cpaelzer> I'd ask for an explicit "could not find consistent pool for pool_file" or such
[12:49] <cpaelzer> 68: yes I see what it is doing
[12:50] <cpaelzer> 61+68: both suffer from the implicit RC issue
[12:50] <cpaelzer> they are the last commands and are meant to return the RC to the function above
[12:50] <cpaelzer> but, due to the set -e if they do so it will abort right?
[12:50] <cpaelzer> and if one adds another command in the function afterwards it will change behavior
[12:51] <cpaelzer> I'd consider it better to collect and return the RC explicitly to avoid the latter issue
[12:51] <cpaelzer> for the former I think it is fine as it is meant to break on an issue
[12:51] <cpaelzer> OTOH if you want to have all tests run in all cases and know which of all of them fail
[12:51] <cpaelzer> instead of breaking on the first
[12:51] <cpaelzer> then it would need changes
[12:51] <cpaelzer> rbasak: is that what you wanted to know?
[13:21] <rbasak> cpaelzer: yes thanks
[13:21] <rbasak> They will abort because of the set -e too, and I was missing that before.
[13:21] <rbasak> But only because of the way the function is invoked
[13:22] <rbasak> If it was invoked with, say "if ! foo; then echo "tehre was a problem"; fi", then the set -e will stop working.
[13:22] <rbasak> At that point I think we'd be relying on the implicit rc.
[14:46] <DevNull1> I'm looking to move an existing install of wordpress on apache2 to nginx.  I've got Let's Encrypt installed too.  Is it as easy as removing apache2, installing/config Nginx, redoing .htaccess and then re-doing Let's encrypt?
[14:46] <DevNull1> It's just a stock Wordpress install - nothing fancy
[14:49] <kiokoman> idk but i don't think it would be that easy, nginx does not support .htaccess
[14:54] <DevNull1> kiokoman, Yes, I am aware. It will have to go in the vhost config
[14:55] <DevNull1> Anyone have http2 working on Ubuntu server 18.04?
[14:55] <DevNull1> Can't get it working and I've followed all the docs -- no dice
[15:05] <tomreyn> DevNull1: there's #nginx if nginx functionality isn't working properly. but make sure your client supports it properly, too.
[15:06] <tomreyn> DevNull1: you migration apache httpd -> nginx may also run into issues about switching the php model / API. there is no mod_php for nginx, in case you were using this so far. so you'd need to switch to fpm (which is better anyways IMO).
[15:07] <tomreyn> that's only relevant if you weren't already using fpm with apache httpd, of course
[15:20] <DevNull1> tomreyn, Thanks I'm gonna rebuild from scratch. Take this as a learning experience. =)
[15:21] <tomreyn> good luck
[16:58] <compdoc> if you have a 32bit ubuntu system, is it possible to upgrade to 64but using apt-get? or do you need to wipe?
[16:59] <compdoc> *bit
[17:01] <_KaszpiR_> compdoc reinstall
[17:02] <tomreyn> it's not an 'upgrade', it's rather a replacement.
[17:04] <compdoc> sure, if you reinstall
[18:08] <nacc> rbasak: i think you can drop our local oauth part with the bump of launchpadlib?
[18:09] <nacc> rbasak: and at the same time remove line 194
[18:09] <nacc> (from the yaml)
[18:09] <nacc> i think that will fix the test failure
[21:40] <nacc> rbasak: http://paste.ubuntu.com/p/838jdvRhPN/
[21:40] <nacc> is my theory
[21:40] <nacc> i am struggling to get snapcraft to actually run right now
[21:42] <nacc> stgraber: trying to see why my `lxc launch ubuntu:x` doesn't seem to be doing anything, and passed -v, still just get "Creating the container" and then silence...
[21:48] <nacc> stgraber: rebooting seems to have made it well again; oh well
[21:49] <rbasak> nacc: thanks!
[21:49] <rbasak> I'd just about got as far as being able to reproduce locally before I EODd
[21:49] <rbasak> I couldn't find where launchpadlib actually used oath
[21:49] <rbasak> But if we can drop the part then it doesn't matter I guess
[21:53] <nacc> rbasak: i'm doing a cleanbuild now locally, and will run a selftest
[21:53] <nacc> see if i can get it to best
[21:56] <nacc> wow, brain failure, *to pass
[21:57] <nacc> rbasak: yeah, the version of launchpadlib in our snap *before* the commit in the above paste did you use oauth
[21:57] <nacc> but now it does not
[22:01] <rbasak> Ah, that makes sense
[22:02] <nacc> so it's really a consequence of the bump, that we also needed to shift this part of the snap
[22:02] <nacc> I probably won't have an answer until Monday locally. I might just push up a MP so that CI tests it faster
[22:04] <nacc> https://code.launchpad.net/~nacc/usd-importer/+git/usd-importer/+merge/345963
[23:31] <nacc> rbasak: ok, that fixed the unit tests, but failed integration tests, which i just pushed a second fix for (bumping lazr.restfulclient to not need oauth by moving to 0.14.0)
[23:32] <nacc> rbasak: i'm thinking we need to drop our strict versioning in setup.py actually, and maybe express minimums (if we know they are needed) or just use the latest?
[23:38] <arooni> what server does cron use by defualt
[23:38] <arooni> grr... shell
[23:38] <nacc> arooni: probably /bin/sh ? If you need a specific shell, you should invoke it by path
[23:40] <arooni> nacc: can i change the shell per command
[23:40] <arooni> or do i need to set it for the entire cron set of tasks
[23:41] <nacc> arooni: i'd check `man 5 crontab`
[23:41] <nacc> arooni: it appears to be a global in the crontab
[23:46] <sarnold> /bin/sh -c "/bin/tcsh -whatever tcshs 'programs use'"
[23:47] <sarnold> probably far better is to just write what you want in a shell script and execute that from crontab. don't get too clever in the crontab itself.
[23:50] <nacc> sarnold: good point