=== Serge is now known as hallyn
=== shokohsc62 is now known as shokohsc6
[11:55] <Lope> can someone please help me figure out what's going on?
[11:56] <Lope> I've installed Ubuntu Server 22.04 minimal inside a KVM virtual machine with it's NIC bridged to my local LAN bridge.
[11:56] <Lope> It has an IP, I can ping it.
[11:56] <Lope> It has SSH installed, I've been SSH'ing into it and using it.
[11:56] <Lope> Now it doesn't want to accept my SSH connections... I didn't change anything
[11:56] <Lope> nc -z ip port; echo $? is returning 1
[11:56] <Lope> it replies to my pings.
[11:57] <Lope> if I run `netstat -lntup` I can see sshd listening there.
[11:57] <Lope> I haven't installed UFW or iptables.
[11:57] <Lope> I don't have any firewall rules on host preventing me from SSH'ing into it.
[11:57] <Lope> I didn't change anything. I just can't SSH in now for some unknown reason.
[11:58] <Lope> Also bizarre, when I let a ping to the VM run, I'm getting very unexpected horrible variance in ping times.
[11:58] <Lope> 0.7ms 24ms 0.6ms 56ms 78ms 101ms 2ms
[11:58] <Lope> WTF?
[11:59] <Lope> CPU is Ryzen 5900x with performance governor activated. CPU is idle. I haven't done anything weird or new. It's suddenly just acting the fool.
[12:00] <Lope> If I run htop inside the VM, the vm is idle.
[12:00] <Lope> The VM has 4G RAM 3.4G available.
[12:01] <Lope> It's a normal KVM VM made by virt-manager running on Debian Bullseye with 6.0.0 kernel
[12:04] <arraybolt3> Lope: You don't have any firewall on either the host or guest, just to be clear?
[12:04] <Lope> I have some iptables rules on the host.
[12:04] <arraybolt3> But those used to work.
[12:04] <Lope> nothing weird, just masqerade etc.
[12:04] <Lope> Yes and I haven't touched them
[12:05] <arraybolt3> What's the exact error it fails with when trying to SSH in?
[12:05] <arraybolt3> Also, is there any chance that maybe your system is under DDoS attack? That might explain the irregular pings and inaccessibility.
[12:05] <arraybolt3> You might check the incoming bandwidth and see if someone's messing with you.
[12:05] <Lope> I've got a debian VM running bridged to the same LAN, and the pings to that VM are all under 0.1 ms
[12:06] <arraybolt3> Yeah that's odd. Is it possible the Ubuntu VM is running out of memory?
[12:06] <Lope> I don't see how this VM could be under a DoS attack, I haven't forwarded anything to it.
[12:06] <arraybolt3> Maybe it's thrashing, or under high CPU load, or something.
[12:06] <Lope> Like I said the host and the guest are idle.
[12:06] <arraybolt3> Meh, you just said it has RAM available.
[12:06] <arraybolt3> Sorry, re-reading now.
[12:07] <Lope> either way, both are idle.
[12:07] <arraybolt3> OK. What's the exact SSH error you get? Or does it just hang?
[12:07] <Lope> host shows 40 Gi RAM available
[12:07] <Lope> It's not really even the SSH I'm concerned with at this stage.
[12:07] <Lope> I can't even connect to the port with nc
[12:07] <Lope> and pings are taking >100ms to complete
[12:08] <Lope> I'm just shocked.
[12:08] <Lope> I've just done a fresh install of 22.04
[12:08] <Lope> did some basic stuff, installing net-tools dnsutils htop locales less. Basic stuff.
[12:08] <arraybolt3> Weird. If it's not under load or attack, then that's pretty much exhausted my (rather limited) ideas. I'm not really a networking expert so it's unsurprising I don't know what's going on, but... uh... I don't know what's going on.
[12:09] <Lope> It's so weird, I don't know.
[12:09] <arraybolt3> Others in here might (read: are likely to) have better ideas.
[12:09] <Lope> oh crap. looks like I have an IP conflict
[12:10] <Lope> I've shut the VM down and I'm still getting ping replies
[12:10] <Lope> hahahahha
[12:10] <arraybolt3> :P
[12:10] <Lope> XD
[12:10] <arraybolt3> Well at least that explains it :D
[12:10] <Lope> thank god. I was going nuts.
[12:11] <Lope> there was a time when I didn't have DHCP setup, and I manually assigned IP's to phones etc.
[12:11] <Lope> So now I've got the fun task of auditing every device.
[12:11]  * arraybolt3 should learn about networking at some point - this stuff both fascinates and scares me
[12:13] <arraybolt3> Anyway, glad to at least have been a somewhat-more-responsive rubber duck for debuggging.
[12:13]  * arraybolt3 has to go afk
[13:45] <athos> ahasenack: I wrote a long reply on that log4cplus symbols file MP at https://code.launchpad.net/~athos-ribeiro/ubuntu/+source/log4cplus/+git/log4cplus/+merge/437047
[13:45] <ahasenack> ok
[13:46] <athos> tl;dr: we are missing the "-fvisibility=hidden" build flag for the cmake builds. If you agree with the analysis, I will proceed to add it and re-generate the symbols file
=== chris15 is now known as chris14
=== Eickmeyer is now known as NoLongerEickmeye
=== NoLongerEickmeye is now known as NotEickmeyer
[18:38] <evit> Anyone know if there will be Ubuntu Pro officially on any other cloud providers other than the big 3
[18:39] <evit> What if I don't use AWS, Azure or GCP
[18:49] <jchittum> You can still use Pro via Token, purchased directly from Canonical
[18:50] <evit> Yes, but I have to pay a LOT to be able to do that vs. monthly usage pricing
[18:50] <evit> There is no virtual instance pricing at all anymore
[18:51] <evit> Only SERVER pricing like its 1999
[18:51] <evit> The big 3 cloud providers include the instance and the PRO subscription together right?
[18:52] <Eickmeyer> evit: Seeing your previous questions in here, you really need to get in touch with the sales team, they might be able to customize something for you. This page, scroll to the bottom: https://ubuntu.com/pricing/pro
[18:53] <jchittum> So a few things: 1. the pro on the public clouds does _not_ include support. so it's closest "price" is the $500. 2. definitely use the "Contact Us" to talk about VMs. The form really only does physical side. as Eickmeyer said, definitely get in touch with the sales folks
[18:53] <evit> They said it would be the same price for a physical server
[18:54] <Eickmeyer> If you've already spoken to them, there's your answer. Nobody in here can override that.
[18:54] <jchittum> Huh...i'm actually shooting off a message now to check that. feels weird
[18:55] <Eickmeyer> Does feel weird tough, as jchittum said.
[18:56] <jchittum> i know there are steep discounts for "bulk" purchasing of tokens. it does get hard, since tokens lack the metering provided by the cloud marketplaces
[18:59] <evit> Yes, I only need 2-5 so I don't think I will qualify. 
[19:00] <Eickmeyer> Did you actually talk to the sales team?
[19:03] <evit> Yes, they were even confused and said I would get a discount and then said it would be $500/server/year for my 2 -- 1 CPU, 16GB tiny cloud instances. =(
[19:06] <Eickmeyer> Well, here's the thing: you already got the information you were seeking. Ubuntu Pro being available on other cloud platforms isn't up to the server team, so the only people that *might* know anything is the sales people.
[19:08] <ravage> the pro site does not actually exclude any cloud services. but nobody seems to know how to actually manage them
[19:08] <ravage> until they sort out what they actually support i keep my local repo cache for my proxmox containers :)
[19:09] <evit> ravage, Yes, I understand that but I think they should rethink the pricing or offer a virtual instance listed for purchase. But hey, I'm just a customer or potentially a former customer.
[19:09] <ravage> im totally with you
[19:10] <ravage> the gap between free and pay a little fortune is really strange
[19:10] <ravage> *I
[19:10] <evit> Also, I don't WANT support. I just want patches for universe and those should eventually flow back to the community
[19:13] <ravage> again preaching to the choir here :)
[19:15] <evit> Looks like Canonical is setting the stage for acquisition and healthy revenues will help the spaceman himself buy another trip but this time to the moon.
[19:15] <evit> =p
[19:21] <jchittum> i really hope that's not the case. i _like_ working at Canonical :)
[19:21] <jchittum> evit: if you only need 2-5, you can do that for FREE
[19:22] <jchittum> personal use supports a up to 5 for free
[19:22] <evit> jchittum, Thanks I am aware of that
[19:26] <esj> fwiw, pro physical server is $500/yr, which comes to about 5 cents an hour. looking at pro pricing on aws, that's roughly the xlarge instance types level - bigger than that and, if you leave them on, full pricing is cheaper than on-aws. smaller than that and yes indeed you're paying more by buying those tokens. but the pricing is really quite similar unless you're running nano's or micro's ?
[19:31] <evit> esj, fwiw that is a HUGE price increase, did 0's an 1's suffer from inflationary pressures?
[19:32] <evit> esj, a small cloud instance shouldn't be charged at the same level as a full server
[19:33] <esj> evit: you're exactly correct; ideally there would be some sort of instance pricing for such cases, and we'd love to bring pro to other clouds. jchittum has raised this internally to our folks, but as mentioned here even with your low volume i'd suggest you get in touch with our sales and see what they can do
[19:36] <evit> esj, Will do. I'm still a bit annoyed but hey life is only a series of endless headaches
[19:37] <esj> yes i'm sorry; i promise there are no ill intentions here, it's more of a pricing gap, but i realize that's little comfort to you who are stuck in this situation
[19:41] <evit> esj, well I guess execs need Yacht's 
[21:06] <bancroft> what would make my eth0 and tun0 devices both have the same inet? and how can I tell tun0 to use something else?
[21:39] <sdeziel> bancroft: what do you mean by "inet"? IPv4 address?
[23:13] <bancroft> sdeziel: yes, it's next to ifconfig so I figured it would be more proper to call it inet 
[23:13] <bancroft> *next to it in 
[23:19] <sdeziel> bancroft: can you elaborate a bit more on to why you'd like the same IP on both devices?
[23:19] <sdeziel> well, maybe a high level description of what you are trying to achieve