up_the_irons changed the topic of #arpnetworks to: VPS Services: https://arpnetworks.com/vps | KB: https://arpnetworks.com/help | Twitter/IG/etc: @arpnetworks | @up_the_irons, @mercutio are staff, other ops are long-time customers | Channel Log: https://irclogger.arpnetworks.com/arpnetworks/
<up_the_irons> Happy Friday y'all!
<mhoran> Happy Friday!
<mhoran> Almost the weekend, time for updating my ARP Networks VMs!
<up_the_irons> :)
<todd> on one hand I have a client needing updating to do dkim for google mail reputation reasons, and on the other I've got a DR failover for another client tomorrow from one region in azure to another. I keep thinking about my arpnetworks vm's, they are my precious, they keep working, I rarely exercise their full potential, someday .. ;-)
<todd> though it is nice to know OpenBSD has wg(4) and I can use it with egre(4) to provide myself with an arpnetworks v6 autoconf'ed address from remote. sssssh, don't tell Gerry he might object... oh wait, did I say that out loud?
<todd> (roadwarrior type remote, my starlink gives me nice /64's for each subnet, though i think dhcpcd is limited in its thinking power, after 16 it gets unstable..)
<up_the_irons> hey I didn't know OpenBSD had wg(4) built-in
<up_the_irons> that's cool
<todd> I had a friend who pointed me at tailscale .. but to do that on OpenBSD required a go userland binary to route things through tun .. when i realized the tech was in wg(4) I just set it up directly. much improved rtt.
<todd> so if you don't mind me asking, when you fought and won with gmail.com .. were you at full block before recovery or .. any tips? .. I've implemented dkim and even that isn't enough yet. I'm still holding out hope that postmaster tools will notice dmarc and things will start recovering from there.
<todd> Not to say you're using starlink, but if someone were, it's fun to use wg(4) over v6 to a remote system with a static IP and then route traffic over that for remote IPSec vpn's that require a static on my end. ;-) glad to share if useful.
<todd> oh dang, I just realized I chose the wrong remote host for a static ip. rtt is a useful metric. provider 1) (hetzner vm in us), 2) (arpnetworks vm in ca)..
<todd> v6 hop v4 hop total
<todd> 1. 85.927 ms 72.544 ms 158.471 ms
<todd> 2. 63.517 ms 63.618 ms 127.135 ms
<todd> total difference: 31.336 ms
<todd> in reality I won't notice. but hats off to your connectivity once again ;-)
<up_the_irons> Oh we never had a block in gmail; we never sent 5000+ messages a day. But in general, setting up dmarc/dkim/etc. was kind of a pain
<up_the_irons> happy to provide good connectivity :)
<mhoran> Latency across the country to my VPS is better than the AWS VMs I have to use for work. By a noticeable margin, since I use my VPS interactively for weechat and mutt.
<mhoran> But I wouldn't complain if ARP were to open an East datacenter. :D
<up_the_irons> hehe
<todd> I'm all hands on if OKC were chosen. Probably cheaper to get a decent business starlink uplink than hard wired connectivity around these parts though.
<todd> hah. a local datacenter (fullnet.net) pings (at least their website host 'fullnet.net') 8ms slower from my vm at arpnetworks than my starlink router here locally. 38 vs 30 ms. heh.
<todd> it would be some serious work but the first hosting company to rideshare a starthink satellite to space is gonna get some attention. I am certain datacenters in space will be a thing. how big a thing, dunno. but .. ;-)