sarnold | nicoz: you know it :D I'm always preaching the "a low end vm in a datacenter is pretty cheap and reliable" gospel to any who will listen :) | 00:38 |
---|---|---|
Unit193 | sarnold: https://www.vultr.com/pricing/ those micro ones. :3 | 02:35 |
sarnold | Unit193: yes! :D | 02:36 |
sarnold | wow, $14k PER MONTH machines.. | 02:37 |
sarnold | https://www.cpubenchmark.net/cpu.php?cpu=Intel+Xeon+Gold+6342+%40+2.80GHz&id=4485 | 02:37 |
sarnold | okay, those are decent procs.. | 02:37 |
Unit193 | If you want dedicated, I'd go OVH. | 02:38 |
Unit193 | (https://www.ovhcloud.com/en/bare-metal/prices/#filterType=range_element&filterValue=high%20grade) | 02:40 |
Unit193 | https://eco.ovhcloud.com/en/ huh, Germany and Poland now too eh? I knew about the latter, but neat. | 02:41 |
Unit193 | ...OK I've gotten side tracked. | 02:41 |
sarnold | $372/mo for their silliest :) heh | 02:45 |
leftyfb | I prefer containers on my own physical in-house servers :) | 02:58 |
sarnold | i still haven't gotten around to containers yet | 02:59 |
sarnold | some day | 02:59 |
leftyfb | Just picked up another R620 which will be a mirror for the one running everything at the moment. I'll do periodic snapshots of the containers nightly | 02:59 |
leftyfb | love me some lxd | 02:59 |
leftyfb | yesterday setup a full lab environment to test out some deployment code. Got a router and "metal" server providing DHCP and DNS and containers on that "metal" container for the deployments just like we do in the field | 03:00 |
leftyfb | lxc launch ubuntu:22.04 new-server | 03:01 |
leftyfb | that's it | 03:01 |
leftyfb | then mess with lxd bridges and profiles if you need to do the fancy stuff I was doing | 03:01 |
leftyfb | lxc list # to list your servers and lxc shell <server> # to bootstrap it with ssh and such. OR do what I do and use ansible to create, bootstrap and deploy everything to it in a matter of minutes | 03:03 |
sarnold | that's so cool | 03:04 |
Unit193 | Never used lxd, but I have lxc for autopkgtests. | 03:05 |
leftyfb | lxd is better for managing multiple containers and allows more orchestration | 03:05 |
leftyfb | you can also create and manage kvm/qemu vm's with lxd | 03:06 |
Unit193 | Oh dang, it's actually in the repos. Huh. | 03:07 |
leftyfb | I just set that up for our support engineers. They're all still running 18.04 because they need to support customers on older versions. But I wrote an ansible playbook to spin up the 20.04 VM, copy over all their keys and creds and runs a full engineer deployments on it | 03:07 |
leftyfb | Unit193: lxd is snap only btw | 03:08 |
Unit193 | !info lxd unstable | 03:08 |
ubottu | lxd (5.0.1-2+b1, unstable): Powerful system container and virtual machine manager - daemon. In component main, is optional. Built by lxd. Size 13,185 kB / 46,440 kB. (Only available for linux-any.) | 03:08 |
sarnold | \o/ | 03:08 |
leftyfb | I can't believe you guys haven't messed with lxd. It's trivial and so easy to spin one up for quick tests and then just stop it or delete it if needed | 03:09 |
sarnold | "messed with', a bit, yeah | 03:09 |
leftyfb | it's how I check package versions and such on releases I'm not running all the time to help in #u | 03:09 |
sarnold | well, okay, lxc list just now is taking its sweet time.. I think I've got one for every supported release, and some of them have esm configured, etc | 03:10 |
sarnold | oh i've got debians going too | 03:10 |
leftyfb | oh man, are they all running? | 03:10 |
sarnold | no, all stopped right now | 03:10 |
leftyfb | how many you got? | 03:11 |
sarnold | but they're all so cheap that I often forget I start them, and I might at some point wonder why i've a thousand processes running | 03:11 |
sarnold | 12 | 03:11 |
leftyfb | lol | 03:11 |
leftyfb | I wouldn't call that "haven't gotten around to containers yet" :) | 03:11 |
sarnold | well, some folks really live and breathe containers, organizae their entire workloads in them, orchestrate them together, etc.. | 03:12 |
leftyfb | that's sort of me | 03:12 |
sarnold | I just have a few sitting around for easy question answering, like yours | 03:12 |
leftyfb | though we're going to be moving to ESXi at some point in the near future. Gotta rewrite all my ansible for that next | 03:12 |
sarnold | but no ansibles to create and configure them on a whim, etc | 03:12 |
sarnold | aww ;( | 03:12 |
sarnold | I thought with the broadcom aquisition everyone was moving *away* from esxi | 03:13 |
leftyfb | yeah, they need HA and support | 03:13 |
leftyfb | hm | 03:13 |
leftyfb | didn't know about this | 03:14 |
sarnold | ah, I dunno about lxd's HA, but I wouldn't be shocked if esxi has *better* HA regardless | 03:14 |
sarnold | another 15 years on the market will do that :) | 03:14 |
leftyfb | lxd's sort of have HA but it takes a bit of work | 03:14 |
leftyfb | you need to strap on some other pieces to do the heavy lifting | 03:15 |
sarnold | broadcom has a history, as it were .. https://www.techtarget.com/searchdatacenter/news/252521669/Broadcoms-acquire-and-axe-history-concerns-VMware-users | 03:15 |
leftyfb | welp, I was waiting for "smee" to come back but apparently it take multiple hours to install a single package. I can't wait any longer | 03:18 |
leftyfb | cyas | 03:18 |
sarnold | gnight leftyfb ;) | 03:19 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!