#networking
1 messages ยท Page 41 of 1
And btw that's why I want a Linux mobile smartphone
How lawmakers and the police think technology works: https://www.youtube.com/watch?v=JMIHNiR3CP8
Clip from episode 2 of "Red Dwarf: Back to Earth".
lol
fairphone runs a de-google'd android, thats the closest thing i know
tbh i find degoogled android kinda unstable
I know
But there isn't really any hardware for it, beside old one plus 6 for example or the shitty pinephone
Also you can't have your banking app or other apps that you NEED anyway ... So you still need two smartphones
Unstable? But it's basically identical
my bank app just pops up a warning you have root/jb
The only difference is Play Services not being present (which can be a problem)
I'm talking about Linux mobile, no android osp or whatsoever, "real" linux
nothing limited coz root/jb just a warning popup every few weeks
Oh you can hide that from your apps using Magisk though you shouldn't have to
Imagine if your bank warned you if you were running an OS on your desktop you installed yourself
No thanks
magisk hide useless now
the guy that made it now makes software to detect magisk hide
he changed teams
Realistically any API an Android app can use you can make lie
<put changed teams meme here>
Still if you got Linux mobile you won't even have the apps anyway
You'll need to use waydroid or smth, and I'm pretty sure this comes with lots of problems
If Magisk is no longer relevant, it's still going to be quite possible to make something else do the same thing
one app i dont know why has to detect root or not is wetherspoons
i just wanna order a drink why do you cair i have root
here ya go
Solution 2 is to just refuse to use everyone's shitty apps
I like that one quite a bit
I like it too
just not needed
Not everyone needs to have their dumb app on my phone
Use the website
I don't have all the options
have you seen most banking sites on moble
And yeah they are shitty
Yes
they not at all worth it
like reddits mobile website
reddit just keeps showing them dam popups to use there shitty app
Discord went out of their way to ruin the web experience on mobile
yeah but someone seeing discord on my phone isnt as bad as reddit
I will never install the Reddit app
same
On Android I use Reddit mobile website to not have ads
Also I don't want their freaking app
So when I have a popup forcing me to go to the app, I leave the website
Not ever, I can't be convinced
Same
some apps i can unoficaly use via web with a little api messing
like the bus tracker app
Hmm that's interesting
If Reddit blocked mobile and old.reddit fully tomorrow, I would simply just not use it
tryed to use it with microg but dident work so next best plan
reverce the app and make a crappy web ui
They are really close to the point where I stop using it already
I'm seeing everything become shit to use and frankly we all lived before apps and I am happy to do so again even if it involves binning everything that refuses to comply
i dony expect a small bus company to make a very good app
hell they will still give you a cd with app source on it now if you just ask
like you know github exsists lol
Dear Google,
You should be like the small bus companies,
Kind regards
True but there are things you don't have a choice
Like my pingID app for work ...
Work wants an app? Provide a phone to me to run that app
So I hope the way they'll emulate apps on Linux mobile will be stable
My phone doesn't do apps anymore, sorry
the app can also melt your phone just zoom out and load every stop lol
Or I'll run it in the dev simulator
Yeah ...
But then they'll tell you to use the windows pingID app on the laptop they give you
And I honestly don't want to use this shit ever
we need this modded into cyberpunk
ยฏ_(ใ)_/ยฏ
lol
so mutch optmization could happen here with loading every bus stop
but it do be a very very nice way to stress out your phone
Knowing they call me with my personal phone I guess I could ask for one ... But pretty sure they'll won't accept
Me phone doesn't need that to be stressed
lol
It's looking like Android app compatibility's becoming more of a thing for Linux-based phones
mine just turns on its fan to keep cool when i zoom out lol
If they want to insist on the security features or break bin them
wait there are phones with fans?
Yeah a few
ye red magic 3 lol
Oh that's nice
Still need to have at least one recent smartphone supported though ...
They all have awful specs because they don't have volume and the good SoC's don't have open drivers
only has enabled in game mode but made my own rom for android 10 and added temp control to fan
I know and that's the most annoying thing
only things not working now is
rgb strip
game mode switch (idfk what to use it for)
triggers work but no way to bind to anything
other than that and 1 or 20 bugs its fine
one annoying af bug
if you dont end call
phone still thinks its on a call and dont let you change audio volume only call volume
could fix
cba
no one realy calls me anyway ๐ญ
a fun simple app you can make
capture bluetooth data that iphones always transmit
im still rocking my hand-me-down iphone 7(?)
interesting, thanks
fun to go to an apple store and see demo allover lol
can also show who has a icloud locked device
interesting way for a scammer to prove they're ur grandma ig
disabled = icloud locked
off is phone is off
and who tf uses siri is someone useing siri
siri used to be good when u can ask her where u could hide a dead body
true lol
now shes boring
there is also other data thats sent out
like encripted clipboard data
encripted audio streams if on call
wow ok thanks for another excuse to get an android
its just to keep a mac and ios in sync
like calls in mac and stuff
you can jailbrake it to turn them off when you select off
Cool - so you can turn off sending data everywhere if and only if your iPhone is vulnerable to a known exploit capable of persistant or semi-persistant privilege escalation. Cool
and yet you cant sideload apps
Oh you can. You just have to either use a developer key and sign them yourself, or use a version of iOS with a known exploit capable of persistent or semi-persistent privilege escalation.
-As you do
have to resign app every week tho
or every month if your a developer
and limited to 3 apps as user
5 if developer
@clear igloo I am surprised Linus was allowed to do all of that lol
"allowed"
LMFAO
They're more like guidelines ๐
rules are like a nose
only there to be broken
Do what?
Itโs no secret their chairs are great! Check out Secretlab at https://lmg.gg/secretlab
Get 69% off any of XSplitโs video tools. Use code LINUS at https://lmg.gg/XSplit
The free wireless internet we installed for the lttstore.com merch team has been awesome, but because we didnโt install it properly, itโs limited to speed of 1 gigabit. Today, w...
Go into their telco's rack & switch, change out transceivers to do simplex Internet from their ISP switch to their CPE & use the other fiber to go from their switch to another switch for building to building
"hey lets let linus come and swap a cable what could go wrong"
linus: drops all the servers
^ Anybody in Ontario know wtf somebody wanted with my Bell box
I found it open just now and the open seal was on the ground like 10m away
I'm not on DSL or fibe but does the rogers coax go thru there as well?
I looked inside. One of the coaxes are cut. Everything still works though. Any thoughts on wtf this could mean appreciated.
I mean either way I'll call Bell but it's gonna be a btch because I'm not their customer.
BGP
But you're going to need to pump those numbers up
Le'groan.... BiDi is great for SOME things... CWDM is a way easier way to overcome lack of fiber
26 routes is NEVER enough
More accurately, any x number of routes will eventually be not enough
I'll just remove everything and send everything out to wan /s
I'm not a network pro so BGP is black magic to me. I just know enough to know that when someone says "X number is more than we will ever need" say NO
BiDi is always preferred if possible and in budget. Very common in SP space
Those are nice ๐
Although multimode only though, great for in rack gear or short haul if your plant has MM
Setup ExaBGP and see how how much of full tables their end can handle 
lmfao
bgp is just my friend announcing his routes to me and vice versa
bc we got lazy & tired of static routes
and it required both of us to update our routers every time we wanted to change a subnet or add one
Anyone else remember when TCAM hit 512K routes in 2014?
IPv6 is pushing us to the next breaking point at the current rate
I heard you need more /127s ๐
And now a nexus 9300-fx/fx2 is verified up to 471,500 yet at that time, that many crashed the internet
You should see the S1 9800 ๐
I think our MX10k3s are rated for 80m/7m in RIB/FIB and the latest Trios pushing a tad bit more
In our isolated network we use 9504s as our routers so I have no idea what those do but we don't really push them
Depends on the linecard, the 9500s with EX hold less routes than the FX
@clear igloo ๐ญ ITS removed Make Me Admin rights

I can't get local admin on our broadcast machines anymore
i'm looking to get a Mikrotik CRS317-1G-16S+ but i'm not familiar with any of the 3rd party websites that they're sold on. does anybody have recommendations for a store to purchase it from?
Everything I've seen about those seem a LITTLE bit sketchy
Not like red flag level but don't expect Cisco/juniper/dell/ubiquiti quality
mtik are fine, just routerOS is not that polished
I bought my hex off amazon
yea the OS seems basic but really just need something basic with a decent amount of 10G ports
Amazon? so brave, i wouldn't risk doing that ๐
I finally installed a Intel X520-DA1 in my server today, and went to run 10G SFP+ modules on both my router and server. But for whatever reason the router did not seem to like the 10G module, it just was refusing to properly negotiate. Router is running an Intel X520-DA2 for context.
Had to run 1G Cisco SFP+ instead and its working now. Not sure if its just a driver issue on OpenWRT?
should we pull both cables? no there no reason too. ๐คฃ
Itโs no secret their chairs are great! Check out Secretlab at https://lmg.gg/secretlab
Get 69% off any of XSplitโs video tools. Use code LINUS at https://lmg.gg/XSplit
The free wireless internet we installed for the lttstore.com merch team has been awesome, but because we didnโt install it properly, itโs limited to speed of 1 gigabit. Today, w...
You forgot to also use 127.0.0.0/8 for something
I am still sort of at a loss, with SWAG, anyone available to help? I know I asked yesterday but hoping someone is out there
I have it set up, ports are forwarded just need help implementing a proxy-pass to jellyfin, it seems to be HTTPS only, so not sure how to get there. I can access swag remotely from WAN, but when I enter my subdomain it defaults to the SWAG correct setup default page.
Rather than being jellyfin.example.com it just defaults to example.com
Same if I do pihole as well. Im missing something somewhere
have you watched any videos from linuxserver or space invador one about setting it up and the config files? or asked in the linuxserver discord?
isn't it literally the same issue they shown in that LTT video linked above, about error corrections being different, and/or about some devices not liking certain SFP+ modules ?
Did you know with a bridge adapter you can share connections to multiple routers, LAN, wifi at the same time? I found it today.
are ugreen's ethernet cables okay?
i've had a few of their products since anker shit itself and they've done exactly what they've said on the tin, just wondering if there are any known issues with their ethernet cables in particular
and i'd rather go with a brand i've actually heard of than the slew of amazon specials
i mean it's probably not hard to fuck up an ethernet cable but i've been surprised before
and since the ltt sponsorships i figured they were okay
and their charging stuff has been flawless for me
cables2go, monoprice, infinite cables, or cable matters are my go to
Do you need preterminated or bulk cables?
oh, i'm not looking at fibre optic stuff
just a straight up standard ethernet cable lol
I know, but you can get bulk ethernet cable or pre-terminated
IE do you just want like a 7 foot cable or do you need 200+ feet ๐
all good, if you just need a set lenght and aren't making your own then yah any of those brands are fine
The issue with bulk (ie rolls of cable) is a lot of companies sell CCA (copper clad aluminum/aluminium) cables which are junk since it's much more brittle/prone to breaking and cannot be used with power over ethernet
hm!
thank you
also, how come i see companies like ugreen advertising cat7, even cat8 '40Gbps' stuff when the more reputable, less 'out-there' companies you've mentioned do not?
is it just a matter of 'weh, we'll just slap this on it'?
yes
Cat7 doesn't have any electrical standard to adhere too and technically to be "in spec" needs a not RJ45 end
Cat8 is a real spec but nothing exists to do 25Gbps or 40Gbps over copper RJ45
hm
The reason is 10G copper uses a LOT more power than fiber. I think a 10G fiber optic is like 0.1 watts of power compared to a copper port at 2w and a 10g copper SFP+ module is like 2.5 to 5 watts
Mikrotik has a list of approved resellers on their website, I'd buy directly from one of those in your region for support reasons. Here in Australia I use Duxtel, wisp.net or Streakwave depending on price and availability https://mikrotik.com/buy/
They're very reliable in my experience, I've only seen two dead out of hundreds in 5 years - one was installed sideways in a muddy puddle by a customer and the wireless transmit stopped working in it, the other I broke the bootloader on while screwing around
I'd rather go mik over ubi any day ๐
Yeah likewise.
Would I necessarily buy one as a core switch handling billions of dollars of transactions? Perhaps not, though redundant architecture matters more than vendor anyway when there is big dollars to be made or lost
Yah, for SMB or home Mik is great, or WISP too
They get used a lot in small ISPs
I have a habit of building product-like solutions around them and am outgrowing their OS a bit that way, but couldn't have even started with UniFi
haha, yah, unifi is great for set/forget and easy setups
If to you a router is a router and an AP is an AP unifi is fine until your iPhones and Surfaces inexplicably refuse to connect
yup
Do you think for short runs it's better to use fiber or direct attach cables ?
Like inside a rack
DAC, easy
Unless you have fiber already run or swap things often
Ok, even though it has higher power consumption?
And why if I swap things often
Oh yeah SFP+ copper module isn't the same as a dac
yah, DACs are a pita to unplug a bunch and I've had plenty of instances with the tab breaking off
Humm that's concerning
And beside high power consumption there is nothing else ?
Price I guess
DACs are crazy cheap yah, I can get a 400Gb DAC for like $100 compared to several hundred+ for two QSFP-DD modules
DAC > Fiber > twisted pair copper in terms of power usage and heat
Fiber > DAC > twisted pair in terms of ease to work with inside a rack (imo) most of the time
DAC > twisted pair > Fiber in terms of cost (usually at 10G and below)
Is there any compatibility issues with DACs like with transceivers ?
Oh and I have (still in shipping) a mikrotik switch with 2 SFP+
I should use mikrotik transceivers ? And for the network card I'll probably buy a used mellanox connect-x3, is there any difficulty choosing the transceiver ?
Mik is usually pretty agnostic for transceivers and DACs are usually more flexible
For NIC cards outside of Intel branded ones I've not had issues with specific ones not working from various vendors when it comes to SFP modules
What would cause a laptop to randomly just turn on and off the wifi
stupid pfsense2.7 fresh install,
can't install packages because it cannot retrieve package data or so.
dns is working absolutely fine, it can resolve internet hostnames, traceroute also works fine when I trace 1.1.1.1 for example
Anyone here use or know of TCPoptimizer and do u know if its safe?
not needed
Used it LONG ago, not needed these days
gehhfhj?
All of their 12G converters are SINGLE MODE ONLY
oh, lol
all of our other fiber stuff is multimode
rip
Do they offer compatible optics?
yea
nvm actually
the compatible optics are just single mode
rip
In regards to DAC cables, Intel are fantastic. At least when using Cisco cables in my experience. They just work every time.
Wait, what are you media converting there? Some make sense like HDMI but are some of those just form factor conversation?
I think it is just transmitting the 12G-SDI over a fiber instead of copper
The internet either isn't helping or my question is phrased wrong.
If I have 5gig lan but my motherboard is capped at 2.5, would a 5gig lan card improve my speed or would it still be capped at 2.5?
5gig lan card ? Did you mean 10Gig ?
I was trying to get help to choose my hardware for my server in the tech chat channel but people keeps telling me that the cpus i'm looking at are bad for gaming or whatever lmao
So i'm asking here
I see some e5 v4 14c/28t for around 25eur on ebay, or 300eur with the x99 motherboard+32gig of memory
Do you guys think it's a good idea ? If not what do you recommend ?
SDI which typically goes over coaxial but this makes it go over fiber
I don't understand, you are recommending me the same ones as the ones that you are telling me to not buy ....?
I was trying to buy the cpu + a compatible motherboar, not go for an off the shelf refurbished server
I don't know, you can just type e5 2680 v4 on ebay for example you'll see a bunch of those CPUs sold like thatt
Oh yeah true
But aren't those super noisy ? I worked with dell poweredge servers and I don't want this in my garage lmao. Even if i guess i can always change the case/cooling/fans
Also I don't like proprietary thingies like idrac and co
Yeah okay
And i guess you don't interact that much with it anyway
If you want efficient and quiet consider spending the money on a desktop motherboard and CPU, modern cores go way further
yeah but then a modern desktop mb and cpu are either more expensive or less performant for the same price
300EUR might get you an old server motherboard but it'll also buy you a new one and more
Performance of cores of that generation is significantly less than a modern desktop CPU per-core
Per core yes
So yes you get less cores. But your cores might do double the work.
Overall a 2680v4 cost 25eur and has the performance of a r5 5600X while having more pcie lanes
Well you said you're paying 300EUR for a motherboard
Older used RAM is tons cheaper
So be it
Lenovo servers are fine as long as you don't buy absolute minimum spec
I remember one genius I worked with who specced a server with no NICs
that is an awesome website
Dude no offence but Dell and HP aren't be-all-end-all for everyone and being overly dismissive isn't constructive.
They're overpriced as fuck for one
I've refurbished dell and lenovo servers and they felt similar build wise etc
That said buying rack servers for home use sucks
They're louder
Less expandable
All those PCIe lanes you spoke of often aren't reachable in 2U
Well refurbished isn't the right wording, more like the company was the one that manages the logistic of "demo servers for future clients", so when they came back from potentail clients we had to make sure everything was okay + reset them
Yes exactly what I thought
That's why those atx chinese x99 motherboards that go for 100eur are so tempting
Is there a lab gopher for europe
Do you think before you talk? 2U uses risers for PCIe, and has limited slots
often only 1 or 2 U, often there is not a lot of slots in the case etc
Most only expose either half-height PCIe or only 2 full slots
Servers with more PCIe slots in that size are a fairly new thing
yeah or use strange 90degree adapters
That's the riser
Yep, they exist
But most won't give you that many unless you pay extra for it
Additionally those cards are size constrained over a tower
That and your rack server is likely quite new
Yeah OK so it's a 2U chassis explicitly designed to have a lot of PCIe expandability
Though I will say it's not that awful, if still constrained
They covered it up by hiding a USB NIC behind the server and not telling anyone why
"Oh we encountered a driver problem"
God knows why Lenovo lets you spec a server with only a BMC NIC and none for the server itself lol
dell will let you spec without a NIC as well now
Oh fun
mainly the OCP ones
wait nvm, there's still 2 Gigabit ports on the board.
I thought those were from the OCP.
And why is it that hard to find the same kind of refurbished servers for epyc ?
epyc is relativity new
Epyc's still quite new
all of the cheap under $1k servers you are finding on ebay are Xeon v2/v3, maybe v4
xeon v4 is just one year younger
Same reason why you aren't finding Xeon scalable for cheap either
Expected server lifespans are getting longer at the same time too
And then you have the people going cloud/hybrid
ziply
Hello, recently my isp switched me to fiber and everything is fine, but they changed my router and i cant access its settings page(it doesnt load). The ip is on the back of the router and it matches my default gateway when i look it up(192.168.100.1). I restarted my router and my pc a couple of times and tried accessing it through my phone and that doesnt work either. My friend told me to call them to ask for access, but im trying to avoid that. ๐ can anyone offer some suggestions or help?
Have you tried carrying a laptop and an ethernet cable to where the ISP's router is, plug directly into the back of it (most have 4 client ports), and access it from there?
i havent, but can try that
My isp only provides a fiberswitch for fiber.
Just to confirm this is an actual router?
Some CGNAT do 192.168.. ,Some do 10...* And some do 172.16.. -172.31..
its a router/modem pretty sure, my pc (tho i had to get a basic switch to extend the cable that goes to my pc cuz i cba pulling cables through walls) and the tvbox are plugged into it and nothing else
its some huawei unit, says its ip is 192.168.100.1 on the back of it and so does my default gateway when i checked it, but doesnt load(says it times out every time i try it)
Try importing the security certificate in your browser from the router ip, and then try using both http://192.168.100.1 and https://192.168.100.1
i think i tried that before, but i tried again just now, it times out with just http, but with htttps its just unable to connect
nvm give me a bit
i cant find a way to import the router security certificate to my browser wihtout having access to my router page, but my browser isnt in https only mode so not sure if that will help
any recommendations on a router i should get, i need at least wifi 6 at 1GB but i feel like 6e would be more future proof
So you want a router/AP combo
pretty much
Do you want mesh or a standalone
itโs just that it has a coax modem so i was wondering if i got an adapter to use coax if it would work like a router
either or, it doesnโt really matter
Can you passthrough your modem to a router
I guess your current ISP provided modem is a modem/router/AP combo
no, so they provided me with just a modem. along with the router they provide for free. I have this old router/modem/ap combo that i want to use with the others, probably as a mesh tho
this is the modem i have
and i want to try and use this as the AP/router
Top list imo:
Netgear RAXE300 $340 (if you need more streams/coverage go up to the RAXE500 $550)
Deco XE75 Pro $300 + Upgradeable if you need more coverage and 2.5g lan
Ubiquiti Dream Router $200 (Great ecosystem & management, only wifi 6 not 6e)
Netgear's Orbi series is very good for a mesh system but insanely expensive compared to Deco...
but would this modem/router work with my other modem if i got the correct adapter?
The netgear you sent?
yes
Yeah it should work, where is the modem going into currently
the modem is currently going into this router
It would be Wall modem coax > netgear router > switch or whatever else you plan to do
And you want to remove that router altogether
the thing with this is that there is no coax that comes out of modem, itโs just ethernet cause we have fiber now
Wait so you have fiber into your home?
yeah
Why would you want to go fiber to coax
I don't believe you would be able to use that netgear router then... As it doesn't have any WAN ports
I want it to go from the modem ethernet to an coax adapter into the netgear router and have the router as my AP
The only thing I know of is a moca adaptor, which is intended to be used from the router into the coax ports in your home
I don't believe it can be used as an ethernet > coax > router
But I may be wrong
yeah, thatโs what i originally thought i just didnโt want the old guy to go to waste
is there a sticker with the model number on it? they might host the web interface on a nonstandard port
does the router matter that much
Not with AT&T
AT&T requires their gateway with their ONT
As the gateway does authentication
I have AT&T fiber, and switched my gateway into "transparent mode" so my own firewall system has the IPv4 address directly on it.
the problem with IPv6 is they use 6-over-4 gateways, and while they do issue me a /60 address, they don't route the whole thing to my firewall, so I'm stuck on a single /64
Incorrect-ish
They actually split the /60 into 2x /61s so you can get 8x /64s and they hold 8x /64s for some stupid reason but only allow you to request in /64 increments
Depending on your router you can request multiple /64s from that /61
Usually :xxx0 through :xxx7 are reserved and then 8 through f are handed out per /64 PD request
If you have OPNSense or pfSense then you can use this method:
https://github.com/lilchancep/att-pfsense-ipv6
There are similar options for Unifi too iirc
@clear igloo Can this stupid company stop using public IP ranges for internal NAT
never!!!
I had multiple customers that I had to explain that just because it starts with 172. does not mean it's private
I had not read that tidbit before. Thanks. I'll give it a try
Yep. A customer's internal network started at 172/8 and subnetted throughout their network. Well they were attempting to move some of their services to the cloud and lo and behold, a lot of that cloud providers space was in 172. After escalating to me (because they blamed us), attitude change once I pointed it out over a vendor meet.
I could hear the anger in the silence. I only needed to see part or their routing table but it was at least a few hundred /24s that they had to re-engineer
that's what happens when u pull a stupid
Just pick up a WAG-D20 and spoof your AT&T gateway and get rid of that thing, true ip passthrough!
- get SFP+ & 10gig
That's classic. I worked for a company that widely used what I consider "Illegal IP ranges", because they didn't own all of them. Of course, they could never interact with the actual owners, but mostly they stuck with US and UK militaries ranges, and they weren't customers.
They used them as if they were all private IPs, all internally. Still, it was simply poor IP planning that required that. They didn't NEED a /24 in every location
@clear igloo Dad: "Why can't I print?"
Also dad:
typically, what style of antenna will be better, a sharkfin or the generic looking pencil shaped antennas?
i googled one of the nubers at the back which turned out to be the model number and apparently i control it through my isps website...
and im given access to limited settings
@hollow marlin WHAT
MY ISP is doing INSIDE ONTs NOW???
And thatโs a way smaller ONT than what I have
Looks like a Gigaspire
Previous job I worked with Adtran 5k chassis, but never their ONTs. I hear their solid and cheap. We deploy Calix and have been deploying Gigaspires too but their license cost are quite high
I want my parents to cut isp voice
And either just ditch home phone entirely or go full voip at least
so I can rip the voice cat5e line out
If the voice line is already fed from the ONT, it's already SIP (99% of the time anyway)
If the ISP is friendly enough, its a simple config change then lol
I will say that I honestly have no clue what this is in that pic
I mean, its not a big deal since the power goes inside anyways to my UPS but I'd love for it to be inside
? Maybe their XGS-PON stuff?
I bet that's how my ISP is doing 2/1G and 10/5G
with my luck since I already have an outdoor box they probably deploy these?
or since its a big change anyways they move it inside? Idk
Can XGS-PON run with GPON down the same fiber?
I mean we deploy ONTs inside without the housing all the time
GIMME
I'd love to just shove that into a switch and carry my WAN over a VLAN into my router
In theory sure because they are different wavelengths. But that would be a mess
oh
that makes sense yea
Tbh I don't even trust my ISP to give me 10gig to anywhere on the internet lol
We can guarantee you the bandwidth, but I am still waiting on the customers that NEEDED 10G to even break a gig lol
I believe we have XGS-PON in the pipeline for orders but not sure when we are getting them. Primarily AE right now
AE would be cool but not practical
Yep, that's why it's being considered now. We're scaling too fast and cannot run the fiber fast enough
isp here:
yea just leave bare fiber spool next to the ont
(not mine bc i dont have fiber but same country)
I would be surprised if they can hold a solid connection
who?
My ISP? lol
oh
that would kinda explain why steam downloads are fast
Home -> ISP network -> (5) ISP peer IRIS Networks -> (6) Telx Colo/IX -> Steam
because steam can max my gigabit connection
The photo of the bare fiber lol
my adsl be like
loose connection for 2 seconds just to mess with gaming
Having worked with ADSL, that's just a feature
Well, "hand out" lol
Fios deployed it then blackholed routing a day later. I still get DHCPv6-PD but cannot route outside Verizon
rip
I don't get anything with this :(
packet cap doesn't show any RAs or anything
also i know that they have main fiber line 150 meters from my property yet the provider claims that it doesent know about it
You technically don't need RAs for DHCPv6 to work. Did you try requesting a /64 and see if you get any hits?
Yep, but BCP is to always enable it
What if you ping FF02::1 on that interface?
RIP
yup...
When I asked them a year ago At this time we do not support IPv6, however this is something that will be supported in the future.
the way my ISP is setup, they are under another ISP that is doing IPv6
cant you just switch to that isp that is doing ipv6?
No
sounds like a monopoly then
The electric coop owns fiber, they lease fiber to the ISP. Which is a dba with the main company being the other company
here most of the infrastructure is operated by national telecom
but any isp can leese that infrastructure and still have competitive pricing
or they can lay their own infrastructure
also that's not really what a monopoly is
I have other ISPs I can choose from to get services from
And the electric coop still owns the fiber, they can choose to not lease to the current ISP
Who owns the poles plays a larger role in fiber deployment than people realize as well
here their pricing is
5.97 eur/month from distribution box to customer
6.60 eur/ month for 1u rack (passive optics)
The electric coop owns them :D
They own literally every pole in the area including all of the ones near homes
And backhaul too
Along with all of the underground
electric coop needs to start a isp themselfs
The only issue like that, ownership, theyโve ran into is the biggest HOA
No poles allowed in the HOA, all underground required
I wish I had your networking problems
Thatโs why they did fiber and lease out
Being an actual ISP is a lot of work for them
Yep and most likely a hell of a lot cheaper per month than paying licensing. I believe we pay somewhere in the $17-25 per pole/month
That they donโt wanna do
i should do fiber and leese said fiber 
Do you have the capital and rights for poles/whatever
Damn and that adds up
neighbors prolly wouldnt be against it
"but why does my internet cost $100/m?!?!!"...well that is a large reason lol
ISP underpriced Spectrum & AT&T
Both of who get pole space from the electric coop lol
i will run fiber this summer to my garage as a test anyway
i will use mechanical splices
Last I heard is our cost per customer is around $30/m. That's just infrastructure, electric, licensing, fees, etc
Splicing is simple as long as you have a decent unit. If you do, do not touch a bare fiber BTW
i dont have fusion splicer lmao
for fiber equipment all i have is fiber strippers and fiber cleaver
Apparently we have a splicer at work???
We have all of the shit to do LC ends and spools and spools of OS2 and OM3 & 4
meanwhile 2eur/pcs
or 1e for connector
Although all of that fiber stuff is mainly for installs we do
Im surprised they have one. You need to test it out
All of our โtouringโ fiber isnโt made by us
We buy it from camplex or lanshack
@hollow marlin https://www.bhphotovideo.com/c/product/1030157-REG/camplex_hf_troc4m_0328_opticalcon_quad_multimode_fiber.html
Buy Camplex opticalCON QUAD Multimode Fiber Tactical Cable Reel (328 ft) featuring Four Fiber Broadcast Interconnect Cable, Heavy Duty Reel with Side Flange, opticalCON QUAD NO4MX Connectors, COCA Factory Terminated, GR-326-CORE Endface Standards, Neutrik opticalCON ADVANCED System, Tested for < -55dB Return Loss, Diagnostics Documentation Provi...
Itโs so expensive for what it is lol
tactical
what is it used for?
Events, concerts, festivals, etc
with those 12g media converters you posted a day ago?
No, this one is for our main network
I thought it was full of marketing buzz words, but turns out it a thing in the broadcast space lol
But yeah thats expensive for internal use
Our current fiber that we are replacing is just 100m of fiber not on a reel with ST ends
The opticalcon is just like XLR but instead of 3 connectors itโs 2/4/MPO fiber lol
They also have โetherconโ which is similar but RJ45
I hate ST connectors, not as much as FC but still. I can see it for this use case though
I HATE ST lol
Itโs not cheap either like itโs baddd lol
We have a few COs with a couple ST panels and a couple FC. FC is worse because it's like ST but threaded and you get .00001mm of space to work with
This is just a fancy Ethercon to RJ45 coupler that mounts in a Neutrik D Series panel
Nooo thatโs awful
Damn, at $12 each, your infrastructure will cost more than the equipment
how many times did the voices tell you to break it to excuse a replacement that doesent suck
Tell me about it
LOL
Pretty damn close when one of the few times I had to deal with it and I couldn't get the damn lock loosened lol
The only thing Ethercon has going for me is the lock isnโt relying on the flimsy plastic latch
Itโs metal so in the field where itโs being constantly plugged and unplugged it wonโt break
I assume that the cable will give far before the connector does
Yea
But then it leads to stupid stuff like this
Where you have a long ethercon run but need to actually plug it into something without Ethercon
So you gotta use a barrel LOL
Which is like fucking $20 bullshit
lol that new Adtran ONT is almost comparable in size
I wanna move more of our stuff to IP though. I really love IP stuff
But the majority of people at work donโt understand IP & networking well
Usually ends with the people that understand the legacy stuff age out or the legacy stuff becomes too expensive
I will soon be finally dealing IP video as our merger completed and parent company serves IPTV. Finally may actually make use of my MC knowledge
Our only issue with IP stuff is just reliability
Most stuff is fine but youโve got the stuff that isnโt soโฆlol.
Hello, I can my client pc can ping the host pc but the host pc cannot ping client PC. Any idea what it could be? I cannot get a tracert output as well.
Probably your a firewall on your client PC/network is blocking it.
Yeah thought it might be that. Many thanks
@rocky badge found in the wild
lol
In my parents' home they have multiple coax connections in two different rooms, one of those connections in the living room is connected to our cable modem and the other ones are unused. We don't use cable TV. I went to the basement where the coax cables seem to all go, and it seems like they are all connected to several splitters and amplifiers (there are other people living in other floors of the house, it's not just my parents'). We want to have a moca link from the living room to the other room that also has coax (they are pretty far apart). Could we just disconnect the cables that belong to these two coax ports from the splitters and connect them directly together to make it work with moca?
I don't see why not
Ok good, thanks
Is there some way to find out which cables are which without potentially unplugging our neighbour's internet? The cables aren't labeled very well.
not that I know of, ik there are ways when the cable isn't plugged into anything but live not an easy way at least
there may be a tool for that, but the normal way is to plug into a device and it makes a tone, problem is that iirc that would interfere with everyone else's connection
@clear igloo @hollow marlin Dorm network provider just took like...3 minutes to hand out a DHCP lease
That's fine right? ๐
noooo
It's a shared network, got to pass the token around the ring
I'm glad TR was phased out by the time I got in this career
Heya folks, I'm redoing my home network setup and decided to decently partition clients and servers on my network, with vlans if necessary.
I wanted to use the 10.0.0.0/8 range, but some folks tell me that's not the best idea in terms of broadcast message and potential vpn conflicts but I'm not sure if those would actually pose an issue? Does anyone have any insight on this?
@clear igloo@hollow marlin Old HS.
Tracing route to one.one.one.one [1.1.1.1]
over a maximum of 30 hops:
1 <1 ms <1 ms <1 ms 172.28.104.1 (Theatre Router)
2 1 ms 1 ms 1 ms 96.4.x.x (School district's /24)
3 1 ms 1 ms 1 ms 172.22.6.201
4 4 ms 3 ms 3 ms 172.22.64.45
5 * * * Request timed out.
6 2 ms 2 ms 2 ms 100.127.64.73
7 3 ms 9 ms 3 ms static-74-218-4-96.unallocated.tn.ena.net [96.4.218.74]
8 9 ms 8 ms 8 ms be2-2902.bb02.atlaga01.bb.ena.net [207.191.187.102]
9 13 ms 8 ms 8 ms te0-0-0-9.bb01.atlaga01.bb.ena.net [207.191.191.158]
10 16 ms 17 ms 17 ms 198.32.132.136
11 8 ms 11 ms 12 ms 172.71.28.2
12 8 ms 8 ms 8 ms one.one.one.one [1.1.1.1]
Trace complete.
``` damn lol
Using 10.0.0.0/8 is fine for the most part, but don't use 10.0.0.0/8 itself, subnet it properly, most likely using /24 s
Are Brocade Layer 3 Switches vender locked for modules of SFP/SFP+ and QSFP/QSFP+ ports? I didn't know if Amazon generic ones would work
Some people say they had no problems, though I'd confirm for your specific model of switch if it's old
It was the Brocade ICX6610-48P-E. I glossed over the manual but couldnt find anything. That's why I'm asking
Yeah people are talking about using all sorts of third-party modules in them so probably fine
what would happen?
we haven't paid for 4 months to the ISP, the ISP would cutout the connection when you haven't paid for that month.
They replaced the Wi-Fi Router when the old one was broken long time ago.
we moved to another ISP now we have 2 Wi-Fi.
Debt collectors probably
I could do with some networking advice. I'm getting a new fibre connection fitted tomorrow with ยฝ Gigabit. I'm hoping to get a NAS soon too.
The internet comes in the living room downstairs, but our home office is upstairs and this is where the NAS will have to be. We live in an old rental house, no ethernet ducts, and no easy way to lay them without damaging walls or leaving them exposed for my cat to pull on.
Any easy renter friendly ethernet methods I could use? Or is that fancy new wifi 6 worth paying for a expensive access point? Our house is fairly small so WiFi signal shouldn't be an issue.
Pretty common on enterprise gear. Wait to find out that Ciena requires a license to even configure VLANs on their switches
SP gear is nuts sometimes!! lol
Even Cisco doesn't charge for VLANs ๐
I got mine with the license included but that is rough for any other owner
Yep...VLANs? RSTP? SSH? Those are "advanced features of a switch, LICENSE!!!"
Here are the license for say even our 3928s and not included in the base
wow!! Well at least telnet isn't advanced right? ๐
oooof, even static routing?!?!?!
Thankfully we have global keys....prior to that I cannot confirm nor deny that many were telnet only because the license cost exceeded the hardware of a 3903 
That's wild
Outside of that and their terrible CLI, at least they are rock solid. If you can get them that is lol
Yah, stability is key
Juniper is at least for the most part honor based. On commit it will say, "Hey, you need a license for this, but Ill configure it anyway". That is slowly shifting with Flex licensing though which is sad
Yah, Cisco (for now) is mostly honor based as well but of course smart licenses and whatnot ๐
Of course the 8000 series routers are different, can't even do an upgrade if you're not in compliance x.x
That's not to say you wouldn't get in trouble if you did run a ton of stuff in prod without licenses but of course if you're doing that you're probably not paying for TAC either
I did not realize some things are honor based with Cisco!
Yah, the only stuff that requires a license to enable and function is HSEC encrpytion and the like for most things*
I'm looking for a cheap network switch that has at least 2x 10G ports and 6x 2.5G any recommendations? I found this one from qnap but that only has 4x 2.5G https://www.qnap.com/en/product/qsw-2104-2s
The QSW-2104-2S is an easy-to-use unmanaged switch with 2-port 10GbE SFP+ and 4-port 2.5GbE RJ45, allowing you to upgrade your network environment by connecting a wider range of devices with different bandwidth requirements. Featuring a near-silent fanless design and compliance with IEEE 802.3az (Energy Efficient Ethernet, EEE), the QSW-2104-2S ...
this guy has 8x 2.5 and one 10gbe
@tawny moat https://www.servethehome.com/insane-48-port-2-5gbe-2x-25gbe-2x-10gbe-managed-chinese-e-sports-cafe-and-hotel-switch-microchip-micron-sandisk/ a bit more expensive and overkill but also an option
do you want sfp+ ports or rj45?
more then 2x the ports for less then 2x the price so per port its still on budget ๐
oh wow thanks for the recommendation
sfp+ is cheaper but its harder to run
it really depends on cable lenght and ammount
and cabling is more expensive
but nic-s are also cheaper
for really short runs you cant beat dac
for 1-2.5gbe?
for 10 gig
yeah DAC's are great, but also very stiff and not very bendy...
so again it really depends on usecase
I have an issue, i bought a server mobo from ebay and it has a static IP for its BMC set. Its in a range i cannot reach from my home network as its out of the dhcp range
https://www.amazon.com/dp/B0C64X3625 8x 2.5g 1x 10g
https://www.amazon.com/Port-Umanaged-SFP-Compatible-YuanLey/dp/B0C64N2QN7 4x 2.5g 2x 10g
probably uses same switch chip as aliex one
i tried resetting the bmc but that doesnt work
any idea how to get to a BMC ip from a network out of range?
do you know what ip is set?
try pluging in directly to a nic
that mobo > cable> laptop/pc
and try to access that ip then
Connect directly and set your computer range to be within that
and turn of wifi i guess, or can i have 2 networks?
You can have 2
In theory, as long as IPs don't conflict which I don't think will
And set static on Ethernet
thanks guys, Also any recommendations for a bit bigger switch? something with 8x 2.5G and maybe 4x 10G?
i havent seen that particular combo
tipically its 2 10 gig ports and then either gigabit or 2.5g ports
https://www.aliexpress.com/item/1005004869154465.html?gatewayAdapt=glo2nld
this one is alot bigger but still worth looking at
Does anyone know how I would setup a reverse proxy (nginx proxy manager) and a custom domain for a jelly fin server (I have the domain from google and the domain is connected through cloudflare)
Yes
thanks for this, actually worked!
Do you need a ubiquiti router/switch to run ubiq access points?
@hollow marlin I want fortigates at work if it didn't require a safe shutdown lol
Juniper is the same. Sudden power loss is almost always a corrupted OS
:(
Yup, that's freeBSD flexing it's strengths 
Pfsense lab Hello Everyone. I have recently started a pfsense router insite of a virtualbox network. I have gotten dhcp and network traffic working. One thing i wanted to do was test blocking sites just for lab porpuses. I have tried creating a fw rule and using pfblockerng and yet the websites i try to block such as youtube twitch or reddit still pass through. this installation does not have much customization. Does anyone know what I could be missing?
How are you blocking?
I tried using the firewall rules to block out bounced traffic and i have tried pgblockerng and tried blocking categories
but how are you blocking? IP? DNS?
pfblockerng has different ways to block
trying to do dns
make sure your browser isn't using private dns
If you use Chrome, click on Settings->Privacy and Security-> Security then under Advanced, see if you have "Use secure DNS" turned on. Turn it off to use your local DNS server instead of ignoring it.
You may also want to block UDP/53 and TCP/53 outbound from anything but the firewall, to prevent bypassing your DNS filters, but beware, it may break some IoT devices, like Amazon Echo, thermostat, etc.
ยฟDid anyone look over my application?
@hollow marlin @clear igloo I am surprised dorms & the main university network don't peer with each other here lol
or that they don't have any sort of interconnect, traffic destined to the university network from dorms go out to the public internet
Like completely outside their AS?
Yeah dorm and main university are on different AS
Bro dorm or uni traffic with compsci faculty people on it...
Could be the most certain way to find most cancerous traffic. Uni kids age plus enough knowledge to like metasploit or some sht
When I was in res, we had like 3 different wifis per floor. What i did was take over one of the routers for myself lmfao. Everybody else just started using one of the other networks
You cant block next ip of bounced traffic with rules by ip. Thats the whole point. DNS i dont even get what youre trying to do. DNS cant block anything really
My uni just have huge trunks to verizon and spectrum
If youre getting proxied or ddos connections, your only hope is block w rules by port. Inbound and you have say ssh on 22 tcp. Just make rule to drop all and only accept from your ips lol 100% will work
Better to only open a VPN port and only allow ssh connections over the VPN network ๐ If wanna go farther
Another way is by protocol. Smart filters exist which can even analyze traffic and block based on analysis
Well if only youre on vpn. But why. If you know youll always have a static wan ip thats yours and only it will need access to ssh... probably safest possible thing to whitelist only that ip
You know 100% unless somebody has physical access to that lan if there be one or hijacks the access point machines themselves... nobody but you will ever connect with that address
Also allowing all from any match and blacklist ban after is just a bad practice in the abstract sense. Your goal is to 'have certain people allowed to shell' most likely. Your vpn rule would mean in practice 'allow all from this iface or ip mask whatever'
Sure but opens you up to physical attack (unlikely unless you're the like a major company) plus then limits you to only being able to access the network while on that network considering how often I've had to access my machine from not my network
Well yeah but physical attack lets be real lol. Even the vpn solution would probably be no less prone to it in the same ways.
Besides theres always ultimate physical attack by going straight to target endpoint and skipping ssh access lmao.
Still would need private keys with a VPN to get in but as I said unlikely with physical attacks lol. And for the VPN you can still make it so only certain VPN users have access they all get their own static IPs when you set it up so you don't have to give the whole subnet access
If they have to break and enter same thing both remote and target box... might as well not bother and hit the target one :kekw:
True lol
Btw how would you make that rule?
Firewall rules you can just say only allow connections to port 22 from 10.10.0.5 for example
By ip or allow as separate interface which is the vpn's. Your firewall is technically blocking and allowing interfaces and ip ranges that happen to be special blocks. Nobody 'should' be on those in the internet
Only people who could be on it would be people with the VPN keys
Plus Encryption is always a nice bonus with VPNs
You know hamachi's ip range. It's 25.0.0.0/8 pretty sure. Nobody uses them. Hamachi claimed first. But those are valid wan ip technically. Also you cant do that for hamachi because they assign random ips from that range to all users
Doesnt matter which network you joined. Everybody has random 25.x.x.x
On network level, something like firewall cant do anything at all
Isn't that where a VPN would be even more useful tho where your IP is random you VPN into the machine and access whatever it is you want to
VPN is a virtual private network. Usually you want the opposite. It depends on use.
Yeah but we were talking about specifically for ssh which is something you'd want to be private
Exactly. So why even open up to 'anyone who auths onto your vpn'. Even of vpn ip and iface, why not just your ip :kekw:
In a physical sense, Id bet vpn is actually more dangerous than real wan ips to some mininal extent whether you like it or not. Youre always just adding extra things no matter what.
Honestly you dont want even shell access to be private all the time. Minimally you always want access to every individual user to be private because thats their intended use. Even still, linux users or actual os, sure. But for ssh you truly want the very username you auth with for ssh itself
Best security is no connection 
Here security gets pretty real and unforgiving. You usually dont even have anything logged which differentiates clients in real world. Even ip. Case and point even ips and all other factors of network connection arent meant to be used as personal user identification. VPN maybe lol its technically better if you auth with same account as billing and crap on service
Thats the old air gap defense as they call it sometimes :p
Thats not really a true firewall. At least for sure not in the networking sense. It's perhaps one of the only widely enough available layer 1 firewalls :kekw:
Honestly whats sad is today thats not an uncommon problem with wireless. Imagine having modem/router which doesnt allow manual firmare updates. On top of that, you cant disable some antenna power or wifi. At most mac filter. After that, wpa2, idk all you have left
So you dont even have practical means of airgapping which doesnt involve custom firmware update through some bug mostly or opening it up and soldering out antenna somehow.
But youre right. No physical connection isnt just the best security. It's the only 100% guaranteed efficient.
even that can be compromised by human error/ social engineering
example: Stuxnet
Just get rid of the humans easy solution lol
Well i mean social engineering can get pretty dumb in not uncommon practical best cases:
You gotta either convince someone to plug cable back into wall lmao
Or maybe like assume false identity of someone they trust enough to change their mind and plug back in
Did you just watch Secret Invasion lol
Btw change their mind plays a big role too lmfao because convincing anybody who airgapped their machine in first place sounds potentially paranoid af and sweaty as bawlz 
Body snatchers? Like the old or remake? Not sure about 'secret'
just use capslock light to exfiltrate data
Im on phone screen portrait lol
Marvel's Secret Invasion
new series?
yup
i will most likely wait untill they finish season or 2 and then binge
Oh idk. But social engineering is a special real life component of attack. It factors into every attack technically if you consider minimally the target itself being information always ultimately reflects some tiny remaining assumptions about how and what the data reflects irl. Whenever its 'personal data' getting stolen, it means already its got something to do w a human owner irl
So hacking somebody's gmail which youve seen them use for dozen other sites as recovery and not one other inbox: you can assume some bank logon you want will also use that email likely. If usernames are only that email, youve gotten the login itself likely
Just host your own email server 
Thats social engineering though pretty sure. Or credential stuffing. You just assume like 99.69 of people use same password across all websites. So you try to use some massive leaked email-password pairs on diff popular service.
Don't actually do it it sucks
How many hops? Curious is it's just that both are peered already at an IX
Have you ever done it? Youre in a world of pain if you do so... trust me.
For instance, a big portion of email providers even big ones (those are the worst tbh). Imagine your servers getting blocked from like all yahoo.com inboxes
You first and foremost absolutely need rdns record for that server. Then you have to bash your head trying to reach their admins to manually unblock you
Its actually kind of sad though. The web is already has really polarized influence on some things like email. Gmail, microsoft, etc any other small or worse yet personal email server can get blocked from the whole world basically
But in the end its everybodys fault for switching to gmail immediately when it was beta lmfao. I ran like the wind into googles warm embrace. Now everybody uses and hosts entire common form of communication w private corporation lol.
Btw that would mean your own interface and auth to that mailbox. Minimally auth. Like tls and imap or pop3 idk lol. Then you connect and get mail with mail client. Its separate thing from web browser
If you want https web access and login, you gotta also run web server just like for sites and buy or find free mail portal on PHP. Or any service which can go through http. Then get ssl cert because especially emails lol. Its doable all free but such a headache. But you still minimally need reverse dns and your own tld.
Sorry tld lol. 2ld. I mean if own tld maybe that could work too. And youre a boss or government somewhere ๐
That's not much work, there's lots of solutions out there for that, the hardest part is the pain of deliverabily
If you are not 100% up to speed on administration and upkeep of hosting your own mailserver stack, you shouldn't touch it, at all.
Yeah the actual deliverability and even having a physical way to accomplish is bad enough. But i remember gmail was a blessing. A nice chunk of my life was before those giants. A lot more people (perhaps majority) were separated into way more tiny mail servers or personal. We had personal one. But spam was insane. Only decent shot tbh was using public spam filter dbs.
There are other things which turn out to also be entirely physical. Like being on personal server means nobody else there by default to get any statistics from when it comes to spam filter
Today, any leet spamma knows that they kiss minimally entire address goodbye if they mass mail people on google or yahoo (idk yahoo because theyve shown 0 friggs given for basic things before..)
All gmail sees is burst of mail origonating from same ip or domain etc to their boxes with no visible history of correspondence at all. Banned. Whatever was common between all flood messages is what they can just autoban without any monitoring. If you used your own private server then sucks lol. Youve just been banned from all google again. And this time admins will review just like before regardless... only tell you to get lost because it's legit reason honestly
It's a lot more complex than that with gmail. You need to configure your dmarc, google postmaster tools, and a bunch of other stuff though.
We got a whole dozen /24's badrepped by gmail because our ops team misconfigured the dmarc record and we sent an email blast to 10 million subscribers ๐
Legit subscribed email list has to have certain requirements met I remember first hand:
- Anybody who gets them must have opted in
- There has to be visible immediate opt out link accessible by all w no requirements
- Must be legit email in of ifself for sure etc
Oof. But irl Ive experienced personally that if you 100% dot is and cross ts, it can take like 1 report even by somebody and you get banned
Now you have to appeal. Theyll probably lift ban but still doing that every couple months? Eek. And try reaching some of their admins in the first place lmfao. Yahoo I legit had to find it first. Then you wait for email back :/
Once you establish a sender reputation (by either domain(s) or ip(s)) with google (and yahoo), it takes a hell of a lot more than a single report (it is percentage based), and you actually have a human review your appeal. Was pretty trivial to get sorted, used to be WAY worse.
So already youre not using your server for at least couple hrs if you did everything as fast as humanly possible
This is why we have numerous pre-warmed ip pools on our outbound mta's. If one drops rep, it is suspended from the outbound pool, and another is rotated in. Not a standard transactional email setup, obviously ๐
Thats my whole point. I think its futile even for mid-sized corporation servers. Like 100 boxes. For legit newsletters, I think you get banned first time regardless lol. Again, it looked like just autoban based on simultaneous incoming identical message
ya, you have to warm your ip's before you ramp up your send. Sounds like you didn't warm! ๐
Well actually now that you bring this up wouldnt that mean you could as easily get notoriously pre-abused ips. Lol it was over from beginning. But here i think its still always doable. I mean theres the dns record and ip/server pair right? If you buy both domain and internet from same person w bad rep... lol nugunhappen irl
yup. You can get that info via postmater tools, but you have to actually have mail flowing before data will show up there. Doing large-scale (legitimate) bulk mailing properly and consistently is a minefield.
Idk. Ours were really warm for sure up until the burst. Thats the whole thing. We had like long correspondences w gov officials for a lot of states (also lol tbh its also kinda messed some of them had official mailboxes on yahoo or gmail). But then whenever our boss sent out those marketing messages... boom. Banned again on same providers even
And we even got banned by hawaii state gov or something lmfao. They personally contracted us and knew who we were but their private mail server was blocking us from the getgo 
Ya. This is why you do transactional and marketing from different ranges. Sounds like your ip's were fine for transactional, but had not been warmed for marketing blasts.
Hmm. Well still, consider how google and yahoo will never block eachother lol. They are on eachothers' whitelists for sender reputable on whole next level of the meaning
When warming new ip ranges for marketing, we cap at 1k/day per mta per domain, and increase it by 500% every 24h (with automatic back-off if the recipient MTA gives a warn response, etc)
Already you have private mail servers guaranteed to never be 100% uptime. But the leading tech giants are exact opposite. And even an hour of being blocked for email is horrible. It defeats the whole purpose of email addr. Its assumed to be where they can reach you lol
mmhmm. I did say not to host your own email :p
Yeah I was stressing to the other guy who pitched the idea lol. Its mission critical
Btw we actually were hosted on microsoft. But one or two of the robots were on our own smtp which we had for some reason...
If i remember correctly, even just the address and personal domain can become issue. So after getting through from server, your mass emails still easily end up in everyones spam on gmail
yup. Large scale email is an art, been doing it for a bit over a decade now. Hate it, but it's part of our toolkit, so
If anything, it might not take more than two people (probably one person irl with two gmails in that case) to report you.
Well perhaps the real large scale is separate and a bit more hidden. It's the real culprit to me: everybody relied on the giants entirely for their 100% personal comminication
can't block gmail, but gmail can block you..
yup. These days it's Google's sandbox, they just let the rest of us play in it (with caveats)
I had issues with deliverability with protonmail a while back, using fastmail now with no issues
Without getting too political ill just say some things got autospammed which i already knew the next place to look was spam folder
It wasnt even like wait half a minute first. They defs had decent enough internet and host speeds at their scales. On top of that, every single person like myself had to opt in for real. And I imagine whenever it was reported, it oppositely was only as not spam
yeah gmail spam is kinda weird, clicked multiple times for my selfhosted (outbound only) UPS email alerts to not be spam, still in spam
Meanwhile on fastmail, clicked once, works just fine
Well again, I feel like today unless you have literally like a third of all peoples main personal inboxes, youll always have some non-zero chance of running into issues. And I stress how with things such as this, it's a showstopper..
What i dont get is those temp email sites. Dont they have to change 2ld and ip all the time?
Yeah. I feel like they dont send to certain email providers. If its against their policy maybe. Others must not care
Fastmail and others now have this masked mail feature which is a random email you can use for a service, and it's receive only
Because in practice, like think about it: even sender cant guarantee which address they land with sometimes. It prevents you from ever knowing anything about sender except they got your email somehow
So unless you sent that email from temp box, its either fraudster or prank lol. Both kinda amp up seriously with the personal factor though 
Oh right derp. Thats how it works usually. No wonder. Kinda cool that its literally impossible to do temp address sending tho
Yeah but it can still get banned outbound just as well. It would make sense too in practice because the whole point is usually verification or recovery method
And I mean you can create as many subdomains as you want, eventually somebody will see that and ban the entire 2ld you own. Tbh if anybody uses subdomain of one you own, sucks lol. I doubt anybody bans anything less than entire 2ld for this crap.
And aliases idk if it makes any diff. You still have the physical ip and its just different domain idk. Aliases i think technically span entire address but not sure
well it's on a service, they def don't use unique addresses for a domain
Honestly, idk if Im even talking any sense anymore. Still dont even understand fully. Like the email addresses arent exactly even a thing from what i understand. Just that theres an expected domain name following a 'name' for user or inbox i guess. We only allow one @ probably because thats delimiter and theres always 2 tokens
no, I have one domain, I can create multple "users" ex: user1@example.com
Just the domain part cant have characters or format that isnt valid for domain. But what about the rest? I guess we make name also valid chars for domain out of simplicity
Honestly a good place to start would be seeing what happens w like postfix and other popular gnu crap when you make linux user name something invalid for maybe gmail to accept outgoing
Or even smtp servers in general. Because there could be user names on linux or something you should deliver to if its set up that way. But nobody sends mail with some character idk.
Whole point is i think it started from user names. Like user@machinehostname. But the email name before @domain doesnt have to be reflective of any user. Nor does it mean its even separate mailboxes
yeah no
Well thats mail tbh. I remember being so lazy one time that i had php sendmail configured to just send root
I wrote i think root@actualsitename.com and only configured that domain name to be servers alias or something. Same with the user. I think i ended up wildcarding somehow everyone to root mail for actual linux user
I was going to say probably on smtp and was correct from the looks of it. The key is in supporting utf8 encoded headers before anything. The two parts making up the address are referred to as 'domain' and 'local' portions. At least in one part of article
Many email clients now offer some support for Unicode. Some clients will automatically choose between a legacy encoding and Unicode depending on the mail's content, either automatically or when the user requests it.Technical requirements for sending of messages containing non-ASCII characters by email include
encoding of certain header fields (...
It makes sense the last part is like fully qualified domain already or some sht. On top of that, it has to resolve because thats the part they chose to indicate where it goes over network minimally lmao. Problem is that its not uri and entire domain name itself is used to identify a real destination over network. Probably exactly why rdns is needed now.
sorry reread my message and by yeah no I was agreeing that it doesn't even have to be a seperate user/mailbox
well yeah, domain tells it the MX server to send to
building a home network, first time, thoughts?
pretty standard, see no issues at a physical side
Smtp is already shoddy asfk. Imo the 'local' part should implement utf8 the same way they did for domains. So not really implement but make the old characters have an equiv unicode representation
rdns iirc is a method just to verify origin server
not needed for actual delivery over the wire, but to get past spam
I don't have rdns setup for home IP, can receive email from home just fine
Yes. But do you need it when connecting to http server? Ftp? No. And its not because you send things one way either
thanks
no?
which switch will you get?
Im pretty sure youre stuck with rdns because exactly that mail server is on network. Everything it transfers with all else is over network. Network endpoints are identified by ip address. You only resolve domain to ip. Nameservers or even names themselves arent used for the actual networking and connectivity.
I am looking at a "TP-Link TL-SG108PE"
yes ik, but it's not needed strictly
as a basic managed switch it looks fine but i would suggest that you get something like Mikrotik CSS326-24G-2S+RM
ultimately depends on what the needs for the network are
Right but ip is the real thing you need no matter what for transferring over network. It also depends on 'network' you could say but thats in a whole other way. Like reserved blocks tbh you know its network thats separate from wan. 127.0.0.1 you know its not even anything beyond network iface itself
Looks good, so does "โข Port Mirroring
- One to One
- Many to One" allow for mirroring for both ingress/egress to one port? saw it on the specs
yes, that's what a DNS lookup is for
send mail to example.com? lookup MX, send to that IP
just regular dns
Yes. To resolve an ip which at any given point represents exactly one endpoint on network generally. So why do you not need it at all for things like http ftp, rtsp, etc
you do
I don't understand your point
Email addresses are poor standard to this day. Like even just 'hostname' portion doesnt even support all new ones. Imo should be done exactly like they did with hostnames call me crazy lmfao
isn't that heavily implementation specific
Many email clients now offer some support for Unicode. Some clients will automatically choose between a legacy encoding and Unicode depending on the mail's content, either automatically or when the user requests it.Technical requirements for sending of messages containing non-ASCII characters by email include
encoding of certain header fields (...
My point is youre referring to actually the very idea of URL. Theres actual technically valid and invalid url. It splits up first and foremost as protocol and everything else is generally an address. But I think thats actually just URI so far. URLs if I remember correctly are valid URI but not necessarily vice versa. The URL begins to define additionally separate part of address (so the token that isnt proto or their delimiter). First half is either full host name or ip addr. The rest is 'path relative to docroot'. But what is the first part generally and why?
well email is old, it's similar to telnet and ssh in the way addresses are defined
username@computer
It's put generally as the network address. It is actually ip address which tells you where http will do the second t part. Because this transfer protocol is over network. The rest is local always with respect to that ip or hostname just like 'user or inbox' is to the second half in email
yes
Right. My whole point exactly. Emails seem to be based on traditions we had with just mail. Unix mail looks exactly like that tbh still. The @ symbol just became defacto the representation of any user on something... lol
That something is minimally a single machine i guess
it's just how it started, no one changed it, and now we are stuck with it
imo it's perfectly fine
apparently it was Ray Tomlinson who made that system
Yup :/
Its far from fine. Because you have secure mail transfer protocol just like any other transfer protocol on same networks. Only the unique addresses for endpoints are not ip address. Literally save an email entirely to eml file. Header body and all. Not sure if even ip it was delivered from is anywhere in there
?
there's received from
by sloti49n35 (Cyrus 3.9.0-alpha0-531-gfdfa13a06d-fm-20230703.001-gfdfa13a0) with LMTPA;
Wed, 12 Jul 2023 16:28:16 -0400```
Ah ok so thats actually in header? With ip?
yes
yes, copied and pasted
Also that looks like rdns isnt set up pretty sure lol. I know its 10 addy but still
actually that's internal email system from
here's one X-Spam-source: IP='54.240.37.196', Host='a37-196.smtp-out.amazonses.com', Country='US', FromHeader='com', MailFrom='com'
Which shows the very thing tbh. You chose to use emails second half for some reason as validated hostname resolving to physical dest server ip. But the transfer itself happened over network from some ip. So you use rdns to find out name
yes and no
Ideally it matches the from address name right? Im not getting it wrong am i?
no
that one was from notifications@stripe.com
rdns is used as one factor in spam scoring and acceptance rules. Determining who it came from is not from rdns at all, but from the "FROM" portion of the header.
there's a From header
^
that's what I'm trying to say
Well i mean in practice idk why you cant just pick yourself the sender to be name@s22-305.east.comcast.50ld.us.com.comcast.com
Theres a diff though
that's the thing, you can put anything in the FROM
You can, but unless you ALSO pass the spam scoring, your email isn't getting through.
at least not in any meaningful way
that's the reason for SPF
SPF, DMARC, and rdns. (plus other things specifically for bulk)
yup
But if email was standardized as valid url just with smtp://samemail.com/userOrInboxWhatev
the format doesn't matter
The addresses not the mailto
it's the underlying protocol
Well again the protocol does pretty much the same thing as http etc. It operates over the exact same layers of network i think. Honestly problem with email could be you rely just on network identity of any sort as persisting information kind of similar to web browsing
SMTP is designed as a Server <-> Server protocol, not really USER <-> Server these days
Only you expect to be replying later to same email which should have the exact same one human as first exchange
you could achieve that with the current format, the problem is again the protocol itself
Wait but smtp is no diff. Its just obscured a bit by how we practically see it. Theres sending mail which is also one way. Theres a difference in I guess the software and even smtp perhaps but only with respect to one another.
If you build an http server into browser, i feel like it would be barely any different
The protocol itself on network though is designed i guess with more than one port for starters lol
But again, I maintain that as protocols, they minimally share transfer over network via tcp at least. Udp actually is actually somehow possible to specify the same exact way potentially. But only implementation would still be different... unless you permit nothing getting across guaranteed
So the to and from have truly at least ip. Forget even making sure hostname anything and all that rdns crap. You have literally cant connect to anything without exactly one ip and port. Net protocols often take care of port separately.
SMTP is over tcp iirc
so idk what you are trying to say
OH!
Irl were assuming that two corresponding sides are on their mail servers. But I think the real problem they both assume they resolve eachothers hosts to same addresses. That it?
15
Yes thats how they fix it. By claiming their own hostnames in the message. Again, as whole email address pretty sure. Why? Lol. Doesnt have anything to do with most protocols involves either. I could be wrong but...
Pay attention to like pop and imap. Isnt username (and its pass) for auth separate from the name@ in email addy
yes, because you are connecting to the IMAP server, not MX
Btw which you get to retrieve mail from same server. Not sending. In fact, is thunderbird or outlook an smtp server? Theres a good way to show what i mean
Mx as in dns record you mean? You dont connect to those at all. Mx ill be honest idk exactly what it does along with most of other record types lmao. Ultimately i know though it is resolved into another name but eventually ip somehow
If a record in dns is another name, you resolve it. Tbh maybe not all. Again, idk what all of them do lmfao. I know cname a and aaaaaaaaa. You also have v4 or v6 address. One or the other. You can fit v4 into v6 but thats not resolution
Mx i think is just smtp server record. Should resolve to ip just like always
When (s)he said MX, (s)he meant MTA, the mail delivery server. An MX record states where incoming mail should go to.
Right. Ideally you want mx to resolve directly to destination no? Either way, still an endpoint correct? The mail travels over network. Again dns confuses me a lot. This is one of those places. I dont even get the correlation between mail exchange and dns. I usually set mx to another hostname like jsbxjxkd.google.com
MX record says "this is the server that should receive email for this domain"
It can point to an IP address, or to an A or CNAME record i.e. ggg.googlemail.com
Right. And what does it ultimately resolve to? How does 'this server' look like?
I'm not sure what you are asking
Lol im too stoops for this. But what happened there? You sent or received outlook message?
dig mx ryois.me Looking up MX record for my domain. Which points to MS hosted Exchange
So any mail going to ryois.me will be sent to that
Idk enough about mail exchange with respect to dns but I always see nothing more than one or two hops on same domain and clearly provider. Is it actually something you want?
Until it gets sent to exchange which is actually the destination server...?
an IP address ultimately
Aha lmao
no that is the destination server
if someone wants to send mail to ryois.me the sender email server looks up the MX record and sends mail to that
Honestly if I had it my way email would be http post and boom. Everything plus more
Then again you would still need server. Otherwise both people have to have their browser + mini httpd running and sender would keep retrying until recipient online and starts theirs
Exchanges i still dont understand because two smtp servers can hold message if anything. Smtp could have just been response code similar to http. Actually lol there is a response code. I remember. Just like http too
Honestly header cant have response on incoming mail except your own. You as destination can be last to give response. I recall response only in failed delivery reports
And it might have been somewhere before dest. I think actually our smtp out might have been the first to get our robot emails. What if mail gets lost somehow in that exchange crap?
once you get an EHLO response from the server you can start sending ยฏ_(ใ)_/ยฏ
if the mailbox doesn't exist the server will return that, storage, etc.
I remember sending to address that doesnt resolve just always gives the quickest delivery notice back.
And once transfer is done, do you always get response with that number status code? Can exchanges also do the same?
did you look at the diagram posted by @rocky badge ?
It wouldn't be less complicated if it were REST, it'd just be happening in HTTP instead/would fail in a single step
You still need all the same components
Pretty moot anyway, ESMTP has been in active use for 3 decades, it isn't going away any time soon
Well if you look at obv things like having transfer protocol of anything and only possible way to find out if/before when the transfer failed... by using itself again for sending that same atom of info i guess backwards 8-)
If that diagram is accurate, my specific beef is with things like send message, terminate somehow, close connection
But again http server I guess doesnt care if transfer successful. Usually client tries same get or load again.
Honestly im starting to realize its not that simple. I mean email as a concept is unique internet comm. Its perhaps all weve ever had for internet equiv of post office and mailbox. Not sure
It expands on concepts that were in use digitally for 3 decades before it....
But i think still like exchanges. Why? Two smtp servers can have mailboxes and send to eachother same way. Honestly client and outbox sufficient unless you want to send message and turn off client
Yes, you clearly want the internet to be a peer to peer architecture instead of a client<->server architecture.
Honestly yeah because smtp daemon would listen and connect
Again, you can have just client send directlt from outbox until it resolves, connects, does protocol crap and transfer finishing with some response OK
Yeah, and if I need to send 10 million marketing emails? Are we tying up my personal system for the hours, days, weeks, or months it will take for it to individually transfer each of those to the individual recipients?
But then its irl equivalent personally trying to put mail into your recipient slot and failing until they open it
Right even that they do irl honestly. Flyers (although lol speaking of deforestation). But the mail is in delivery once you send it. And it ideally remains that way until its delivered
Yeah, that's a non-starter mate. I can send those same 10 million emails, while respecting recipient server preferences, in a few hours right now. It would take a hell of a lot longer with your proposed implementation. Anyway, if you think you have a superior solution, write up an RFC and submit it :p
I need to head to bed, ciao.
Tldr mail over internet or irl is harder. You have mailman too. Client/server? Idk but email aint got it imo.
Like it looks to have been actual user but over network you have no real way of knowing what it is on other end. You still have it because the rest of email only distinguishes dest. But on servers we'd like some local ids unless its 1 user per server
And tbh thats how it all started. We seem to have taken that same mail and began transferring across network with smtp or anything of those sorts. It all started with me noticing that if you have actually personal mail server, you can literally not use the name. Only domain.
Im probably talking sht tho. You need rdns. End.
Quick question here - using VyOS with VRRP to build HA. I only have a small /31, one IP for VyOS and the other one for my firewall. But in the same subnet I'd want to build HA in subnet - essentially using one shared IP in the /31 for VyOS
Does anyone have experience by using VRRP peer IPs from a private network? Does that generally work?
HA, with a client in a /31? Never heard of anyone doing that before, that's generally for point-to-point only and VRRP kinda isn't
I can't say "that won't work" because I don't know, but generally you'd be doing that on a /29 at least, and you can definitely do it on private networks.
I'm aware of that, yes. Usually there're 3 IPs needed - 1 IP per node (2 in total) and 1 Shared IP. In this case, the Shared IP ("being HA'ed") would be the public IP in the /31 and the 2 IPs for keepalive-data-updates would be private IPs instead
On the public-facing eth0 on VyOS I have VRRP configured (I have a /29 there) and it's working fine. But not on the transfer network between VyOS and the firewall. Basically I just don't want waste more IPs then necessary
There's nothing special about "public" or "private-facing" in this context, VRRP only affects the devices within the same subnet anyways
Im beginning to wonder if i might be forced to make the dodgiest networking setup in existence soon
ethernet cables taped to the walls and around doors with switches jammed in dark and forgotten corners type of dodgy
Looks like it may be that way because it's getting scrubbed by Cloudflare. However, paying transport for your own space is $$$
Got my mikrotik switch
So small and light, almost feels like a super cheap device
Which I guess it is 
I know. I'm just concerned the failover not working reliably or causing any edge case issues I can't think of right now.
As it's majorly designed to have all 3 IPs in one subnet, rather than in separate ones
Why can't i access my router login?
how do i make a static IP for a PC?
like 192.168.1.5
Cloudflare?
why isn't the wired device showing up?
in the wired device ipconfig it says the autoconfiged IPV4
why is it auto configed?
i don't want it auto configed
i want it static
so i have put the ip configuration in settings > networking n shit and put it to MANUAL
IP: 192.168.1.5
subnet mask: 255.255.255.0
DNS: 8.8.8.8
DNS 2: 8.8.4.4
I was just looking up the first hop and saw it resolved to CF's space and assumed some rproxying is going on. Of course outbound routing wouldn't be related
; <<>> DiG 9.16.1-Ubuntu <<>> router.etsudorm.ryois.net
;; ANSWER SECTION:
router.etsudorm.ryois.net. 300 IN A 172.67.167.60
router.etsudorm.ryois.net. 300 IN A 104.21.16.102
NetRange: 172.64.0.0 - 172.71.255.255
CIDR: 172.64.0.0/13
NetName: CLOUDFLARENET
OriginAS: AS13335
Organization: Cloudflare, Inc. (CLOUD14)
NetRange: 104.16.0.0 - 104.31.255.255
CIDR: 104.16.0.0/12
NetName: CLOUDFLARENET
OriginAS: AS13335
Organization: Cloudflare, Inc. (CLOUD14)
Oh thatโs my internal dns stuff lol
ryois.net doesnโt have any purpose on the internet, I just have a wildcard pointing to my server though lol
oh ๐
Surely dorm ISP & uni can connect somewhere
There's gotta be old fiber or something from where uni IT used to do dorm networking
nvm i fixed it
does chromium's secure dns interfere with adguard/pihole?
yes
is it okay to turn it off?
sometimes facebook is being resolved on chromium
i have blocked facebook and instagram on my adguard
hi i got a Q? i seen this now some people say that's why you don't have a gig but you do not use pin 4,5,7,8 I don't see why you don't get gig
100Base-T (100mbps) can be over 2 pairs (4 wires) 1000Base-T (gig) requires 4 pairs (8 wire)
thanks for clearing that up๐
Even if not, the cost to run it would have paid for itself 100x over
lol
Yeah in the SP space transit cost is the by far the highest (Cogent, Level3, Arelion, etc). Peering (IX/IXP) is much cheaper. Easily saving $10,000s/month routing internally
Is there something that can test what ping I should get at the moment? I'm tired of hopping in a game and it shows 50 to then start jumping up to 200+ middle of game.
Ahhhh
Yeah uni has DIA from Level3
15 min drive from uni lol
Hi, hopefully some can help me problem solve this. How to stop clients unplugging any 3 of the open ports on a switch and getting a connection to any of the VLANs
port security?
are you untagging at the switches?
Nope, but each switch is going to have to require vlan access for all ap's down chain
so switch 1 is going to need vlan 2,3,4,5,6 access on its chain out
wait, why does every AP need it's own VLAN..
also that network setup with chaining is not really optimal cost wise too
there are reasons for it, that now cant be undone
so working with the above
I understand a dedicated cable back to a managed switch for each ap would be optimal
things I can think of are port security where you tie a port to a MAC address, and just overall physical security

