[00:08] <android808> Evening all.  Question about uninstalling a graphics driver.  I was installing the vulkan driver and now Ubuntu doesn't boot.
[00:37] <android808> I am trying to get my wireless to connect via terminal, but I can't find any posts that work for me.  Anyone know how to do that?
[00:37] <android808> Been trying iwconfig
[00:37] <android808> From posts i have been looking at
[00:37] <android808> And iwlist
[03:03] <cmaloney> jrwren just got told to run stuff on Linux for .NET by Scott Hanselman
[04:44] <_stink_> ..
[13:10] <jrwren> cmaloney: yeah, those .net core folks are such tools. they totally miss the point.
[13:10] <jrwren> it makes me sad how messed up this industry is.
[13:13] <jcastro> jay
[13:13] <jcastro> your tweets frighten me
[13:13] <jcastro> "tip: you won't reboot your cloud insts anyway, right? `apt-mark hold linux-image-generic` to prevent kernel updates from filling your disk."
[13:13] <jcastro> I am so glad you're not on the distro team
[13:15] <jcastro> the rest of your posts seem satisfactory.
[13:19] <cmaloney> heh
[13:26] <jrwren> jcastro: you don't really run ops, do you?
[13:26] <jrwren> jcastro: I don't anymore, but I did.
[13:26] <jrwren> you don't NEED that kernel unless it solves a problem for you, so you don't install it or reboot.
[13:27] <jrwren> There is only a remote vuln fix in a kernel once every few years, so you only NEED to reboot once every few years. The rest is YAGNI.
[13:27] <jrwren> This is how ops is done when actual risk analysis and cost benefit analysis is done before doing things.
[13:27] <jrwren> You do not jsut do things, "because we've always dont it that way."
[13:27] <jrwren> Also, how would being on the distro team matter? Seems like you'd want to keep me out of webops.
[13:28] <cmaloney> I think it depends on the time / effort to spin up a new instance
[13:29] <cmaloney> and if you're going for "long-running" instances rather than just spinning and building new
[13:30] <jrwren> exactly part of why I tweeted what I tweeted.
[13:30] <jrwren> "you won't reboot your cloud insts anyway, right..."  because its short lived.
[13:31] <cmaloney> Yeah, that makes sense
[13:31] <cmaloney> though why you're running apt on short-lived instances also eludes me
[13:32] <jrwren> but also, there is the case of cloud instances which aren't on the internet at all. security group with all inet access disabled. What is the risk of not upgrading kernel on those instances?
[13:32] <jrwren> cmaloney: ha! that is exactly the next step! NEVER upgrade anything on cloud instances, ONLY redeploy to get upgrades.
[13:32] <cmaloney> :)
[13:32] <jrwren> cmaloney: but... unattended-upgrades is on by default and spams your systems automatically with "security" updates.
[13:32] <cmaloney> Yeah, I know
[13:33] <jrwren> but... ya know what those "security" updates do when the root disk fills? IT CAUSES DENIAL OF SERVICE! So you are worse by default!
[13:33] <cmaloney> though Ubuntu Server has the option to not have that on as part of the install
[13:33] <jrwren> cloud-img is on be default and that is what runs everywhere in the cloud :)
[13:34] <jrwren> *in cloudimg it is on by default*
[13:34] <jcastro> and it also prunes kernels by default
[13:34] <jcastro> you have a problem and you went right to "turn off updates"
[13:34] <cmaloney> newer versions, though
[13:35] <cmaloney> I see both sides of this
[13:35] <cmaloney> if you have an insance that's only going to be around for an hour or two then having updates run is pointless
[13:35] <jrwren> it does not prune kernels by deault.
[13:35] <jrwren> I had this happen to me.
[13:35] <jrwren> 1.5G of kernels filling that little 9G root partition.
[13:35] <jrwren> it only prunes OLD kernels.
[13:36] <cmaloney> I've had kernels overflow my /boot partition
[13:36] <jrwren> but if you leave the system *unattended* then it isn't rebooting, it is only installing new kernels which will NEVER be used.
[13:36] <cmaloney> but I'm also on 14.04
[13:36] <jrwren> also, I did NOT "turn off updates" instead, I pinned ONLY the kernel version. Other security updates continue to be applied automatically via unattended-upgrades.
[13:37] <jrwren> as former sysadmins, modern devops requires us to rethink everything, including this. The fact that we aren't shows that we aren't willing to give up the past and change and doing things newer and better.
[13:38] <cmaloney> Newer, yes. Better is still TBD
[13:39] <jrwren> oh, i agree, every use case is different, however, to disregard it entirely is foolish.
[13:39] <jrwren> you'll note I never suggested it should be the default.
[13:39] <cmaloney> :)
[13:41] <jrwren> its also REALLY hard for a lot of us who love tech to admit that newer isn't always better.
[17:59] <jrwren> whoa, community people, anyone know what Joe Landman is up to now? https://scalability.org/2017/03/some-updates-coming-soon/
[19:37] <jrwren> "one of the hardest things when you are trying to affect change is that people like this gentlemen are right, in certain areas."