[00:17] <nacc> stokachu: in the conjure-up snap, why don't you stage the conjure-up part's usr/lib/python3.5? Isn't that where site-packages get installed to? And also since it's after python3.6, wouldn't it be 3.6?
[00:17] <nacc> stokachu: nm on the first, i see it's usr/lib not lib/python
[01:27] <brianw> nice work on lxd integration!
[06:12] <lordievader> Good morning
[08:15] <funabash1> hi huys how can i kill tcp        0      0 127.0.0.1:6379          0.0.0.0:*               LISTEN      5140/redis-server 1 ?
[08:16] <funabash1> kill -9 pid doesnt work it restarts again with a new pid. /etc/init.d/redis-serer doesnt exists
[08:19] <azidhaka> systemctl stop redis(-server) ?
[08:31] <rh10> guys, which approach better and more secure: i setup access via ssh using keys, but anyway i need pass to execute some root command
[08:33] <andol> rh10: Not sure I understand your question fully, but sounds like you might want to configure sudo to allow you to run that specific command without having to enter a password.
[08:35] <rh10> andol, yep. is most properly way to do it in ubuntu using NOPASSWD option in sudoers?
[08:38] <andol> rh10: Yepp, NOPASSWD would be it, and (again) preferably having that sudouers entry be restricted to running a particular command.
[08:39] <rh10> andol, got it, thanks
[08:41] <rh10> another question. is ubuntu docs (official, community, wiki available as packages for install in disconnected networks?
[09:01] <funabash1> guys anyone know a tool for check if your ubuntu server are secured ?
[09:06] <andol> funabash1: How you define secured? Out of the box an Ubuntu server is reasonably secure. Then there are plenty of ways you can create users, install services, etc and make it less secure.
[09:06] <andol> funabash1: On the other hand, if you want it to be secure in a boolean sense there is always the option to both pull the network chord as well as the power chord.
[09:09] <andol> funabash1: Anyway, what tools can help you with is that they can warn you about (obvious) insecurities, but the lack of such discoveries doesn't translate into the server being "secure" per se.
[09:10] <funabash1> andol: like a vulnerability scanner
[09:10] <funabash1> patched version of programs etc
[09:11] <funabash1> lynis maybe is a good tool
[10:53] <nixadmin> Hi all
[10:53] <nixadmin> Currently now server using 99% of my memory, In linux how can i find the exact amount of memory you would need?
[11:14] <adac> ok now be honest who of you is Mark Shuttleworth? :-)
[11:15] <adac> thought this is a nice system montitor, maybe ubuntu likes to add to their packages: https://github.com/aksakalli/gtop
[11:28] <wretchedspirit> adac gtop looks pretty
[11:29] <wretchedspirit> htop is always an option too
[11:29] <wretchedspirit> and it's in the repositories
[11:37] <adac> wretchedspirit, yeap using that one :)
[11:37] <adac> mean htop already. very handy
[11:38] <wretchedspirit> hell yeah
[12:08] <Walex> adac: it is based on 'node.js'!!!!!! :-(
[12:15] <adac> yeah :/
[12:17] <stanfordtools> when I push to git, I'd like my remote server to pull automatically, so that my website is updated. What's the best way to make this automatic? Maybe instead of just pushing I can run another command that pushes and pulls remotely. But how can I make such script?
[12:26] <lordievader> stanfordtools: You mean something like: https://www.digitalocean.com/community/tutorials/how-to-set-up-automatic-deployment-with-git-with-a-vps
[12:37] <stanfordtools> no
[12:37] <stanfordtools> well, kinda
[12:37] <stanfordtools> I don't want a bare repo
[12:38] <lordievader> I don't think that that matters. The point of the tutorial is the post-receive scripts.
[12:53] <stanfordtools> lordievader, I don't need those. I found a way to git pull via ssh
[12:53] <stanfordtools> so, i think that's all i need
[12:54] <lordievader> I guess I misunderstand what you want then.
[20:04] <nacc> dpb1: rbasak: i think i've fixed all the segfaults on x on a, still working through the python dependency chain to make sure the integration tests pass
[20:04] <nacc> rbasak: you mentioned writing a test for the changelog case I hit yesterday (that is, dpkg-parsechangelog skipping some entries). How do I do that in pytest?
[20:06] <rbasak> nacc: drop a changelog file that exhibits the problem, and a test function that asserts it?
[20:07] <rbasak> I don't really know how else to answer the question, sorry.
[20:07] <nacc> rbasak: we would just be testing dpkg-parsechangelog?
[20:07] <nacc> rbasak: so i'm not sure what the test shows us other than you can't rely on the behavior on certain distros
[20:08] <rbasak> nacc: I'm suggesting a test that verifies that the behaviour is the one that we expect, that's all.
[20:08] <rbasak> So the test will instantly fail if using a wrong dpkg-parsechangelog.
[20:08] <nacc> rbasak: right, but I'm trying to understand who uses that test?
[20:08] <nacc> and I don't expect someone runnning from git on xenial to run the unit tests before using the tooling
[20:08] <nacc> since that seems like the 'use case' -- to konw that you can't do that :)
[20:13] <nacc> rbasak: i suppose we could have git ubuntu do self-tests on startup and fail to run unless you're on a 'good' OS
[20:13] <nacc> the snap would somehow avoid that, i thinkn
[20:14] <nacc> at the same time, the fix is probalby to fix dpkg-parsechangelog on xenial :)
[20:28] <rbasak> nacc: that's a good point. I had assumed the test would fail if we do it wrong, but as you point out there's a build time vs. runtime distinction.
[20:28] <nacc> rbasak: yeah, it's messy :)
[20:28] <nacc> implies, to me, that dpkg-dev needs the unit test
[20:28] <rbasak> nacc: perhaps we should run our test suites (or a subset that is possible) after building a snap, from the snap?
[20:28] <nacc> rbasak: yeah, i was going to try that next (via our jenkins job)
[20:29] <nacc> i think i'll need to switch to a shell
[20:29] <nacc> but also means i'll need to snap pytest :)
[20:29] <nacc> so not a priority itself yet
[20:29] <nacc> pytest-3, rather
[20:29] <rbasak> This feels like a common problem for all snappers to me.
[20:29] <rbasak> As-installed testing for snaps.
[20:29] <rbasak> A bit like dep8.
[20:29] <nacc> just asked that i #sanppy
[20:29] <nacc> *snappy
[20:30]  * nacc needs to clean under the n key