<nekobot2>
[haiku/haiku] 455468cbfd55 - sdhci_pci: remove (incomplete) use of MSI
lorglas has joined #haiku
Kokito has joined #haiku
Kokito has quit [Quit: Vision[]: i've been blurred!]
AlaskanEmily has quit [Remote host closed the connection]
grossjonas[m] has left #haiku [#haiku]
acidsys has quit [Server closed connection]
acidsys has joined #haiku
ClaudioM has joined #haiku
bitigchi_2 has quit [Ping timeout: 480 seconds]
vdamewood has quit [Read error: Connection reset by peer]
vdamewood has joined #haiku
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
codysseus[m] has quit [Server closed connection]
Anarchos has joined #haiku
codysseus[m] has joined #haiku
SanzFC has joined #haiku
eroux has joined #haiku
HaikuUser has joined #haiku
lorglas has quit [Quit: Vision[]: i've been blurred!]
bitigchi_2 has joined #haiku
bitigchi_2 has quit [Ping timeout: 480 seconds]
Skipp_OSX has joined #haiku
<kallisti5[m]>
hm.. irc notifications working... looks like emails aren't though. Will investigate
countryboy has joined #haiku
countryboy has left #haiku [#haiku]
Anarchos has quit [Quit: Vision[]: i've been blurred!]
<waddlesplash>
kallisti5[m]: no email from gerrit either
<waddlesplash>
kallisti5[m]: let me know when I can push something to test
<kallisti5[m]>
two issues. the first one was easy to fix (dkim keys, wrong permissions)
<kallisti5[m]>
the second one... eeh. I'm opening a ticket to support
<kallisti5[m]>
tldr; outbound connections come from the k8s nodes directly (instead of from a NAT ip), which means we have three IP addresses that could send mail
<kallisti5[m]>
three IP addresses that could change as kubernetes nodes are recycled :-(
<waddlesplash>
can you just put all 3 in for now?
<waddlesplash>
so we can have email again, it's pretty bad not to get gerrit notifications
<waddlesplash>
esp. because Gerrit has no "notifications center"
bitigchi_2 has joined #haiku
<kallisti5[m]>
that'll be my backup plan
<kallisti5[m]>
actually.. that doesn't seem viable
SanzFC has left #haiku [#haiku]
bitigchi_2 has quit [Ping timeout: 480 seconds]
HaikuUser2 has joined #haiku
HaikuUser3 has joined #haiku
<HaikuUser3>
hey
<HaikuUser2>
yo
<HaikuUser2>
ok bye
HaikuUser2 has quit []
HaikuUser3 has quit []
<zdykstra>
...
<HaikuUser>
:<
Skipp_OSX has quit [Quit: My Mac Pro has gone to sleep. ZZZzzz…]
HaikuUser has quit [Quit: Vision[]: i've been blurred!]
HaikuUser has joined #haiku
eroux has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
bbjimmy has quit [Quit: Vision[]: i've been blurred!]
bitigchi_2 has joined #haiku
bitigchi_2 has quit [Ping timeout: 480 seconds]
<kallisti5[m]>
oof... strike one for Digital Ocean
<kallisti5[m]>
me: Hey, how do we setup reverse PTR records for our kubernetes node pools on Digital Ocean?
<kallisti5[m]>
DO: Why would you run a mail server? We recommend using an outsourced mail server solution. Spam, hard, etc.
<kallisti5[m]>
me: Err, that's not an answer?
<phschafft>
hm.
freddietilley has quit [Ping timeout: 480 seconds]
Vitto has quit [Ping timeout: 480 seconds]
tqh has joined #haiku
bbjimmy has joined #haiku
HaikuUser has quit [Ping timeout: 480 seconds]
<nephele>
kallisti5: just rempoe the spf record or set it to neutral
<kallisti5[m]>
nah.. not even that. I can't assign a reverse ptr record to our outbound ip addresses
<kallisti5[m]>
which pretty much every mail server will reject
<kallisti5[m]>
i'm standing up a tiny dumb vm running docker to run the relay in
<nephele>
why do you need that
<kallisti5[m]>
it sucks... but is the cheapest solution
<kallisti5[m]>
... it's how mail works :-)
<nephele>
I don't see why you need a relay in thos setup, or a ptr record
<kallisti5[m]>
generally mail servers require a valid reverse ptr configured to reduce spammers
<kallisti5[m]>
... to send email?
ablyss[haiku] has quit [Quit: Vision[]: i've been blurred!]
<nephele>
Why are you having three servers in the first place?
ablyss[haiku] has joined #haiku
<nephele>
one is sufficient
<kallisti5[m]>
... ok sorry. I have stuff to do and this conversation isn't helping
freddietilley has joined #haiku
<nephele>
I'm trying to figure out what you are doing, I have obe mail server and many setups that send through it, never did i need a relay
<nephele>
Or do you call the actual mail server the relay?
<kallisti5[m]>
we're moving to a kubernetes cluster with 3 nodes
<kallisti5[m]>
because we're growing exponentially
<nephele>
So you want a mostly identical setup and the mails on each? do they have three dns names?
<nephele>
You could set them up as distinct mail servers habdeling the same domain in dns, if my understanding is correct that should work
bitigchi_2 has joined #haiku
<nephele>
(so mail01. mail02. mail03 MX pointing to all the nodes, and the same for the reverse recorda)
<kallisti5[m]>
once again, none of this is relevant in a managed do k8s cluster
freddietilley has quit [Ping timeout: 480 seconds]
<nephele>
I don't understand why, you said above the ip adressed are cycled, no?
<nephele>
k8s has to build on the normal stuff somehow ;)
<kallisti5[m]>
kubernetes nodes are throw-away
<nephele>
As a bit of motivation though, if you have any setup capable of TLS you are already further than webkits mail setup ;)
<kallisti5[m]>
every node in the cluster can go away
<kallisti5[m]>
and stuff will keep working
KapiX has joined #haiku
<nephele>
that means the new IP adress is arbitraty? If there is no way to know new adressed beforehand a different vm is probably the best setup indeed :/
bitigchi_2 has quit [Ping timeout: 480 seconds]
<kallisti5[m]>
yeah. generally it might work in theory if the reverse ptrs were configured
<kallisti5[m]>
but digital ocean doesn't support this... which kinda screws everything up
freddietilley has joined #haiku
freddietilley has quit [Ping timeout: 480 seconds]