#networking
1 messages · Page 307 of 1
they said it's ethernet
yes
but I mean what port is it connected on etc
because maybe there are two dhcp servers on the internal network, one for the correct network, the other for the wrong one
the second router might be uplinked on one of the LAN ports with the DHCP server enabled or something
I tried bridge mode first, worked fine except Wifi stopped assigning IP adresses, restart etc didnt help so we had no wifi for hours. normal router mode (with static ip adress) works fine until it loses connection to the ISP gateway
what port on the linksys plugs into your ISP gateway
yeah
yeah to which?
1st lan port from gateway to "internet" port of the linksys
as i already said its all hooked up as it should
on both routers sadly, but otherwise wifi wont work
either this or that
the weird thing is we have another linksys router downstairs (older model) for wifi which works just fine in bridge mode
you said you gave your linksys upstairs a static IP address on the internet port?
the static IP you used, that is outside of the DHCP range of the main router I assume?
its literally the ip adress it assigned automatically and i pressed some button to make it static
nothing out of range or anything
idk it says static ip
where does it say static IP - in the linksys or in the main router?
ok, yes some gateways conflate dhcp reservation with static IP
on both?
if the linksys is on a dhcp reservation it wouldn't know that it is
yeah when i set up the linksys router internet connection i chose to type in a static ip instead of using dhcp
maybe thats it
i will change it to dhcp
yes, change it to dhcp
but i can leave it static on the ISP gateway right?
yes
well i will try that, usually it happens <48hrs
when the ISP side says "static" it doesn't actually mean "static", it instead means "DHCP reservation" and assumes the client is on DHCP
it is a way of making sure that a client device will always get the same address via dhcp without actually having to change the device to a static IP configuration
its all in german but it says static ip
i see
the name ending with 0 is upstairs
yeah those are dhcp reservations, if they were exclusions for real static addresses they wouldn't need ot know the MAC address
thats what i set the linksys to
i hope its not bad posting my isps dns and shit public right?
idk
yeah so you can just change the linksys to DHCP
nah those are all private except the isp dns
all we know basically now is your ISP
neutral
wow 😂
neither good or bad, just there
your ISP dns servers are not some secret anyway
oh alright
my pc still has that issue tho
some cheap windows laptop too but its used over wifi always anyways so idc
both have realtek ethernet
sorry what issue?
My PC is connected via ethernet to the linksys router but randomly loses connection
check windows network log
what does the PC show when it loses connection
windows troubleshoot says it doesnt get an ip adress
check the event viewer
not identified - no connection
so you actually know what's happening
kinda
but idk why
wtf theres a lot of errors about the realtek controller and that it has been reset
like every minute
"The network driver detected that its hardware has stopped responding to commands."
try downloading a newer nic driver from realtek
I had a similar issue before.. I would upgrade the driver to the latest one from the realtek site and it would fix it
its all up to date
windows update would download an older driver and downgrade it and break it again
wow
it is? did you go to the realtek site and check there?
typical windows update
i will but as far as i know its updated
i checked like a month ago tho
or 2
if you are just checking for driver updates in device manager it probably won't be the current version
imagine using device manager lol
how were you checking for updates?
i usually use asrocks own shit and iobit driver booster. i prefer downloading it from the actual manufacturer but with shit like ethernet uhhmm im lazy
but theyre usualyl very up to date
but im checking rn of coruse
course
if updating fixes it imma kms
i pressed repair and i think its installing a new driver now
i will report if it happens again
thanks yall!
the asrock website may not be up to date with the latest driver if it is not a new product
if your mainboard is 3 or 4 years old, the driver on the asrock site will likely be out of date, because they don't bother to update the drivers page when the product starts to get older, even though the vendor is still going to be releasing new versions
wireguard is generally pretty fast on everything
asking for full gigabit though is probably a bit much
nice, 700Mbps, I'm getting around 600Mbps lol
with Wireguard
😦
This is on a dual socket Xeon E5-2620 v2
and a 8700K client
Lol
I mean TBH
in most cases that's fine for me
I don't know how wireguard handles stuff exactly, but one issue with tunnels of any type in general is that the performance tends to be impacted severely by jitter (latency variation)
if you have very low latency and jitter, the tunnel throughput will be quite high
Since most of the time...I'll be connecting to it from cellular or public wifi
but out of order packets can make even an unencrypted tunnel operate quite slowly
when I run EoIP tunnels on mikrotik I can get great TCP throughput if the latency is very low, which also means the jitter is low
but when the latency starts to increase to like 50ms or higher, and presumably the jitter is similarly increased, the TCP throughput drops even though the CPU is not maxed out
anybody here running windows 10 pro with hyper-V?
Yes
@little schooner I changed the IPv4 config for my default vswitch and I want to change it back to what it originally was
but unfortunately I didn't take a screenshot of what was there
could you share what you have in your IPv4 settings in the default vswitch config?
@tender hazel
Thanks!
np
Bandwidth delay product (BDP). Higher latency, more time waiting for ACKs, less throughput. Once or twice a month I get escalated a ticket on a customer that is complaining about not being able to reach the speeds they're paying on a L2 circuit for and usually its almost always related to BDP. One of our month engineering training I cover it and its impact
This is how quickly latency impacts speeds
**0ms**
[ ID] Interval Transfer Bandwidth Retr
[ 4] 0.00-10.00 sec 1.13 GBytes 969 Mbits/sec 875 sender
[ 4] 0.00-10.00 sec 1.13 GBytes 968 Mbits/sec receiver
**20ms**
[ ID] Interval Transfer Bandwidth Retr
[ 4] 0.00-10.00 sec 790 MBytes 663 Mbits/sec 743 sender
[ 4] 0.00-10.00 sec 788 MBytes 661 Mbits/sec receiver
**50ms**
[ ID] Interval Transfer Bandwidth Retr
[ 4] 0.00-10.00 sec 547 MBytes 459 Mbits/sec 0 sender
[ 4] 0.00-10.00 sec 547 MBytes 459 Mbits/sec receiver
**50ms** -** 4 Threads**
[SUM] 0.00-10.00 sec 1000 MBytes 839 Mbits/sec 989 sender
[SUM] 0.00-10.00 sec 992 MBytes 832 Mbits/sec receiver```
Jitter surprisingly has far less impact and loss (as long as consistent) is not as much as you think. This means nothing however for UDP of course
BDP is why "long fat pipes" became a thing as TCP and latency is an unavoidable problem
@hollow marlin yes, but I'm not only talking about BDP.. there is an additional hit taken for tunnels
I can do higher TCP rate outside a tunnel vs. inside one, in cases where CPU is not being maxed out
Yeah you have overhead, especially when using TCP as the encapsulation
if a single CPU core is not maxed out, and yet the tunnel limit is quite a bit less than the raw limit, it is due to other factors
some of it is probably mikrotik specific, and some is certainly related to fragmentation
because it is a very large difference - 10-15Mbps max through the tunnel vs 300+ Mbps outside, TCP throughput
so I'm not talking only about overhead from encapsulation
That too, buffers can be several ms depending on windowing and frag. Im just saying majority of the problem is due to latency
yes, in our testing we found we can get very high EoIP tunnel rates where latency is low
but we have one site where we have about 60 or 70ms latency, and I can only get TCP to ramp up to a few Mbps
through the tunnel
outside of the tunnel we can use the entire 40Mbps connection
with a single TCP stream
but inside the tunnel, 1-3 Mbps for the TCP stream
I want to get that latency down so that the users will feel their service is faster
because at the moment they feel it is slow and yet the 40Mbps uplink is usually only about 1/2 to 2/3 used
I would go with it being a Mikrotik thing. Overhead for asymmetric tunnels over the internet is down to the header, frag and latency. Some loss here and there too but not as devastating. I wouldn't expect that much loss unless the CPU was the limit or a core being pinned.
Agreed on latency, low latency is 90% of the experience in most cases
MikroTik EoIP has a feature where you can send 1500 byte packets over the tunnel over layer 2 and it fragments them and recombines them on the other end
we use that to ensure that we can pass any necessary packet size
TCP throughput is higher if we reduce the EoIP tunnel MTU to 1452 or lower
Is this internal or over the internet?
over the internet
You can try prerouting mangle mss clamping to assist if you have not already
yeah it is a bit.. trickier
we are running MPLS over EoIP
we are running VPLS tunnels over that
the VPLS tunnels are carrying PPPoE
so we have a few levels of encapsulation going on
I can try doing additional mss clamping on the underlying PPPoE
I'll just have to figure out how to do it for only those that are going over that link
normally we try to get layer 2 links to sites that are 1550 MTU or larger
if we can only get a 1500 or slightly larger MTU link, or it is layer 3, then we run MPLS over EoIP
Is there any reason why you are running VPLS over EoIP? No need for the extra overhead if you already have a L2 tunnel between the sites unless there are multiple customer behind the Tik at the remote site
yes - on the rest of our network we run VPLS to bring the customers back to the core
so our network is entirely routed, but our customers are brought directly to our core via VPLS tunnels, whether they are getting PPPoE service or DIA
we could run it across the existing EoIP tunnel but the issue is that the EoIP tunnel goes to the wrong router
so we would have to trunk the VLAN across layer 2 over a bunch of core devices
which would make troubleshooting really difficult in the case of a loop or other layer 2 issue
Im well aware of the design, we're all routed with VPLS, any other design for L2 is not sustainable. Im just saying the interface for the tunnel at your edge can be part of the VPLS bridge and be encapsulated with EoIP on egress to the remote site.
Its easy enough to do in Juniper but Id have to dig into what Mikrotik can and cannot do with VPLS
oh yes, that's possible for us to do linked tunnels like that
but I'm also trying to get us to get another transit uplink
and the one I'm looking at has direct routing with the provider that services that site
currently that EoIP traffic takes a boomerang route from winnipeg->toronto->winnipeg
if we get the second transit (which we should have anyway for redundancy), it would decrease that latency to 1-2ms
and then suddenly running VPLS over EoIP doesn't make too much difference
right now we are entirely reliant on hurricane electric as an upstream
we need a second one
Yeah if that path is what its taking, you need another peer. Well like you said, you should have one anyway.
yeah, the problem is that the major local internet providers in Winnipeg (Bell and Shaw) do not peer at the internet exchange in Winnipeg
leading to the boomerang routing unless you buy service from them.. we were looking at buying BGP transit from Teksavvy, which peers with Bell in Winnipeg
and is much cheaper than buying transit from bell, and they give a blackhole community
our old upstream didn't give a blackhole community and so we couldn't have an automated mitigation method for DDoS attacks
we used to have the network taken down for a few minutes every few weeks by DDoS attacks until we moved to HE for an upstream and were able to set up automated DDoS mitigation
We offer both blackhole communities and flowspec for our downstream peers. Later a little more restricted to prevent what HE went through last year but its less impact on our edge as the filters are done in HW and it doesn't require as much TE as its being filtered at the edge instead of the core
Previous SP I worked for had no such desire for it, which we felt when some customers we being hit hard
our satellite network still runs through our partner in Ontario, unfortunately they don't have automated DDoS mitigation and so our satellite sites were hit by several DDoS attacks in the past few weeks, which required their manual intervention to fix
they were fairly quick because at least they now have alerting set up when it happens, but it still takes 5-10 minutes to remedy the situation
5-10 for a DDoS? I figure if you already have solid alerts set up, you can make a playbook for that.
^ Me thinking with my site-based methodologies
On our main network with fastnetmon guarding everything, we block every DDoS attack in seconds
our customers probably don't even notice
but when you are reliant on a human being who has other jobs aside from monitoring DDoS
and they have to log in to devices to create blackhole routes etc.
From an ISP perspective, automation is not needed as blackhole triggering/flowspec is pretty well supported at this point. Simple to implement and not just a benefit for their customer but their entire network
huh, just looked at fastnetmon. Looks like its worth a read of their docs.
Don't think its worth implementing in our hunt kits, but I might yoink how they're doing packet capture
we've been really happy with fastnetmon
we implemented it two years ago on our terrestrial network to send blackhole routes to HE automatically, and we've had no DDoS attacks take things down since then
I can see in the log where it blocks one every few days by blackholing the customer
I mean it may not be the customer's fault that they are being DDoS'ed
but it is a lot better that it is just them going down and not our entire network
Does it have any methods of interaction other than pulling stuff off blacklists? Like filtering of alerts, or redirecting to another endpoint?
how do you mean?
For example, alert is generated from host X dos-ing residential host Y. Before potential blackholing of residential customer, alert is ran past a reputation check, and sees the res IP is "trusted" and only blackholes the sourrce
I may be missing something here, but how do you only blackhole the source?
routing tables in almost every device are based entirely on the destination, not the source
treating one source differently from another is in the realm of PBR (policy-based routing)
Thats where flowspec comes in
So, being honest that I'm overall ignorant of anything that happens at ISP level. Just applying my experience with products like snort that can act as a IPS by blocking IP's that hit on sigs, but additionally being able to be run across a home network filter to prevent a local addressing being blocked
Implemented via a drop at the firewall
ahh I see, yes I am reading about it now. I wasn't familiar with it previously
Frankly the BIG network world terrifies me, where speed is weighted favorably vs monitoring and security
The problem with a firewall drop at the ISP level is that the DDoS attack can consume our entire uplink, so even if we are dropping those packets, it leaves us with no more bandwidth to use
it might be different if we had say 25Gbps or 100Gbps transit
Yeah thats what flowspec does. blackhole triggering (there is a source based flavor) send a BGP update that has a NULL route which essentially takes the customer down. Flowspec does it by sending a BGP update containing any number of src./dst. IP and port numbers and any router using the flowspec NLRI capability with apply a filter on all interfaces based on the BGP update. Much more granularity without the effect of basically just unplugging your router
but with just 10Gbps transit a DDoS attack can overwhelm it, and what we really need to stop the flow of packets is to get the transit to stop sending it to us, not to drop them after the transit has already sent them
We had one mission where we had to monitor a datacenter's 10 gbps uplink. Tuning our sensors to do full PCAP on that was a nightmare.
but flowspec seems like it is intended to solve exactly that situation
Its nice until you are one of the T1s that forgets to add the import policy that denies flowspec NLRI that request to block TCP port 179
if I understand you correctly, you mean forgetting to add a filter that prevents flowspec from blocking the BGP port on your router, which would take down the peering entirely
Customer advertised a route that was missing both src./dst. IP which mean the filter applied to all interfaces blocked BGP for all src./dst. IPs. Peers went down which caused mass convergence, but as things were converging, the flowspec route was obviously purged which removed the filter, peers came up, received the route again....rinse and repeat
Yeah it would be a nightmare situation. 1,000s of routers bouncing routes while the time it took to propagate just cause black holes everywhere
I mean any sort of centralized system can be vulnerable.. and things like RPKI help to some extent, but for the most part everything is based on trust, and everybody being a good netizen
It is, thats why I cringe at the thought of CCNAs that say they know BGP. AKA, here basics how it works and heres how to setup a peer. Not filtering, no security, no design thoughts... Many which of have their hands in public peering everywhere
Its simple but complex at the same type just due to taking in consideration on how it interacts at every point
yup.. it is amazingly simple to just advertise your routes to the peer.. but harder to set up a correct filter to ensure that you don't advertise routes to your peer that you shouldn't be advertising to them
otherwise any random misconfiguration can end up passing up to the upstream in a way that wasn't intended
RPKI is on our to-do list at least. Currently we are using Radb (non-standard precursor to RPKI) but from what I hear from other friends at some T1/T2, they make it sound like its a pain to implement. I know its similar, but one of the few BGP subsets I have yet to even look into
RouterOS v7 has support for RPKI
I don't think they have implemented flowspec support
we had to add a few blackhole routes due to cloudflare's RPKI checker introduced about a year ago
we don't actually implement BGP "safely" as they define it, we just drop the prefixes used by CloudFlare for that test site
we will roll out RPKI when we can, but given that early adopters have run into horrible issues that taken down their entire network in some cases, is it really worth the risk before it stabilizes?
I feel the same based on the horror stories. I know when the time comes I will be labbing it straight for at least a month before I attempt to deploy it. We have tight filtering for our downstream peers on what prefixes we import based on the LOA they sign upon turnup. Combine that with our edge routers that server the sole purpose of BGP to our upstream and their 8-12,000 lines of config, Im confident in what we have...as long as we don't run into many of Juniper's bugs
yeah well that is really where the issue comes in with RPKI, you are placing your trust in the automation doing everything correctly, and when the code is only months or at most a few years old, it is hard to do that
If Mikrotik could add flowspec it would be a nice bonus but its niche for a majority of their customer base. There is also a lot to it in terms of dev as its "automation" in a sense where it can not only filter IPs and port, but also do classification and policing. They'd have to impliment it correctly without interfering with the customers FW or wrecking the CPU
I'm not necessarily too bothered by our customers getting blackholed since it is better than the alternative, and it is more likely than not that they were behaving in such a way online that somebody wanted to attack them
and we have only had residential customers being blackholed, I don't think we've had a single case of that with a business account
Until your own network is attacking itself...remember the Mikrotik bug back in like 2017? Well at the previous SP I worked for we had around ~2000 AC2 in the field and at that time, Mikroik's and their quirks were new to many of the engineers and something along the line of ~800-1000 of them we vulnerable and part of the bot net at the time and had two instances of our own large business customers being attacked internally. The following day is when they released the emergency update and it made the news. I think we had upwards of 50-70gbps of traffic to the core links which was not fun...😆
Due to the design we could not black hole the traffic
Yep, and guess what was not disabled in the Dude to the devices
ugh
mhm
I mean when that vulnerability appeared I was thinking "who would be so stupid as to open winbox and webfig ports on their device to the world in the first place?"
but the answer is a lot of people apparently
I became comfortable with routerOS after that. Spent a few weeks with some engineers learning the ins and outs and from then on out were locked down.
we were locked down before that, so a few managers were concerned about it
but we were like "no, we block those ports globally, so it doesn't impact us"
They had a decent FW setup prior but what I suspect happened is a few were not provisioned correctly and the filters were missing some to block addresses within our space. So 1 could have compromised the rest
Oh we did pretty quick
but I wound up getting a few annoying comments from upper managers like "so this proves mikrotik is crap, we should move to cisco and hp" etc..
thankfully those managers no longer work for us 🙂
Same. But all vendors have exploits. They have pretty on top of patching since
yes I agree - and too many people unfortunately considered webfig and winbox ports to be safe to open to the world at large
generally I only consider ICMP plus VPN protocols to be safe enough to open to the entire internet
Fiber better? 😛
not going outside
Ah!
the only part outside is behind an AP
gotcha
uhmm i just wanted to give an update about the problem we were talking about yesterday, i forgot who it was but ehh the problem still persists. My linksys router still loses connection to my isps gateway randomly
I have a 4u system weighing around 75lbs. every time i've bought hardware it's come with slide rack kits. this time it didn't could someone help me find the right 'quality' tool for the job. I've been told to get an asus r20A 90-S00SP0250T, but they are unavailable in the US
So you need server rails?
What server is it?
it's a comino rm grando
im looking for slide rail
I don't think such exist unless that case has something specific on the side
it does.
Idk if that would fit into those rails
I'm not too familiar with server hardware but it think that would help in a way
@clear igloo lol oops
network ups tools (can't say acronym) shutdown works on everything but esxi
oof
So waiting for everything else to boot up
This cisco switch takes ages to boot lol @clear igloo
My servers boot faster than it
Hmmm, 2960?
Yeah lol
Shouldn't take long unless it's booting a different image compared to last time
Hmm lol
I'm setting up a linux vm for mikrotik netinstall because I am sick of troubleshooting the windows version
and what do you know, it worked right away
Hi, DxDiag says that my pc is Miracast ready but in the setting it's written "This device does not support Miracast reception" in yellow and I can't change any settings below
Btw, this is on my main pc which will be my primary monitor
are there any cheap 10G ethernet switches that are unmanaged and fanless?
same for NIC's
fiber or copper? budget?
There's that mikrotik 10g switch
cheapest 10g copper NICs I know of are around 100, each
But sfp + only
this because i want to run it through my existing wiring, which is CAT5e, but it's not long at all
maybe like 20 meters at best
You would need sfp+ copper modules too
nothing for RJ45?
yes there is a mikrotik 10G copper switch
The copper transceivers are expensive
Because they consume a lot more power
yeah i know that SFP+ modules are expensive, with cheap cards
but i don't have fiber, plus isn't that expensive in and of itself?
wait really?
eh i don't wanna run fiber though
what about 10G fiber transceivers?
14 bucks for 30m https://www.fs.com/products/40206.html?attribute=193&id=99046
30m (98ft) OS2 9/125 LC-LC Singlemode Fiber Patch Cable Duplex 2.0mm PVC(OFNR) at reliable fiber Jumper manufacturer FS.COM. In stock, Custom Service on-line.
20 bucks
You can also get things used
Wait
That's multimode
issue is that i want to connect 2 NAS's to it, one is a cheap Zyxel NAS and the other is a custon NAS
i can easily upgrade the custom one with a PCIe card
but the Zyxel NAS would have to do it with gigabit, which is fine by me, i barely use the thing anyways
but i want it at least connected to the same SAN
what about 2.5/5 then?
That's much more expensive because it's pretty new tech still
dang
10G is the way to go
@late geyser a short distance 10G fiber optic link costs maybe total of 60-70 bucks
if its within the same rack, you can also just get a 10G direct attach cable (SFP+)
those cost like $20
but limited to 3 meters
in a general sense
fair
why not just go short length fiber at that point?
a single direct attach cable costs far less :P
fair enough
you dont need a transceiver or fiber patch cable
but what if you wanted to go for 40 gig later down the road?
I have my server and switch connected with fibers though
40G is a different transceiver, fiber optic (if you use singlemode) is the same
then you'd have to replace the cables and such instead of just the transceiver
@clear igloo let me know a few minutes ago by email that there is an.outdoor AP just above the fire escape door on that side of the building. We should see if that is adequate before adding one. 😩 whyyyyy noooo I don't want to use that shitty AP
@late geyser if you use single mode fiber optics, you can get transceivers from 1G all the way up to 200G
with single mode, fiber itself has no impact on speed
QSFP-DD
but same cable pretty sure
differnet connector though?
but 40G, singlemode LC would work?
actually
QSFP-DD does 400g on sm too and QSFP-28 works in QSFP-DD cages
@tame carbon you lied to me
Cisco QDD-400G-FR4-S Compatible 400GBASE-FR4 QSFP-DD PAM4 1310nm 2km DOM LC SMF Optical Transceiver Module,Product Specification,Product Specification:Cisco Compatible:QDD-400G-FR4-S:Vendor Name:FS,Product Specification:Cisco Compatible:QDD-400G-FR4-S:Vendor Name:FS:Form Factor:QSFP-DD:Max Data Rate:400Gbps
AAAaaaaaa this person is suggesting to use the existing shit AP
I don't want to use the existing AP, its so shit
on a single strand of fiber, no
its only 2.4Ghz, 10/100 Ethernet, and doesn't even work in the configuration
there's Muxing
sm fiber is "unlimited" in theory
or intentionally hamper it
I don't need to
@late geyser you can use a multiplexer
its a shitty tp link eap110 outdoor
then just tell them "it doesn't suffice, we have to replace it"
its already hampered 
just like after one day
keep complaining all day that it has to be replaced
I'm just thinking about reliability and intercompatibility with the parts.
I responded with that
Then they responded with Yep. If we need to add, we should.
Thank god 😩
i would have just replied with "the connection barely reaches and cuts off every once in a while"
The AP they already have isn't even that good
just jam a screwdriver through the top of it
🤣
finally make use of that flathead screwdriver
Caution: not a prybar
I had to install a gigabit switch there today so 
Everything was connected to a 10/100 switch
I still have a couple of those
10m token ring 😄
I have a pile of 48 port 100M switches
Then the gigabit connection going to their router & another switch
I'm also planning on replacing their router too
about that https://blob.rocks/HFGf55kN8F.png
Yeah......
damn
10/100 bottlenecks some shit really badly lol
yeah
We need network connections outside for NDI....
So we're going with this https://blob.rocks/d7pNlVv832.png lol
Mesh pro for WiFi clients, Lite meshing off of the Mesh pro and bridging Ethernet
tested it at home with my stuff
@rocky badge only use I can see for a 100M switch is when you want to do bandwidth throttling without a queue xD
or connecting a bunch of smart home devices, those don't need gigabit
10/100 switches = usually older
yeah but ethernet is ethernet
at my internship i have to work with a 10/100 cisco switch and 3com switch
Plus gigabit switches are cheap so not really any benefit
3com, that's a throwback, next you'll tell me there are some nortel switches too 😛
Lmao
those 48 ports are actually 3coms
just learning stuff dw
pretty sure they use at least gigabit
they have SFP uplink
haha I know, don't need fancy 400Gbit switches to learn on 🙂
hp procurve next /s
not sure what you're talking about because i do
You said you're just learning stuff, I was saying 10/100 switches are fine for learning
certainly not
i need AT LEAST 100G to learn

@clear igloo how do I tell this person the APs don't work with each other
"The AP's are not compatible, now stfu and buy the good stuff"
Aka the APs I'm planning on buying won't connect wirelessly to their existing AP
lmao
Unless they wanna buy ptmp stuff
@rocky badge Just sell them an asus RGB router 😛

goddamnit https://blob.rocks/GShn3n3QU8.png
Yeah the way I've been thinking is cheaper
and can be used as normal WiFi
@clear igloo Is that outdoor AP a TP Link EAP 110 access point? The Ubiquiti UniFi one's wont wirelessly connect to that one. Plus, the EAP 110 is only 2.4GHz which won't be high throughput, which we need for NDI. Ok I'm just telling them lmao
@rocky badge square foot of house is what
?
For what
your house
why
why not
this isn't for home
oh, that field....
imagine not counting hamburgers per footballfield
@thick minnow lol remember avondklokrellen?
yeah
Germans now having their rounds too
with conspiracy wackos
you think a water cannon, pepperspray and batons is going soft?
I mean I have family in germany and they are tired of all the measures
@peak cloak they oughta quit whining :P
apparently they need a specific type of mask?
we dont have a choice dont we
@peak cloak which they can get for free.
it's not that deadly tbh but let's not get political
he was perfectly healthy
I had covid, wasn't bad at all. I had worse
different for everyone
I've been self isolated for a year or so now
I only see 2 other people
who also keep away from rest of society
local farmer's market is usually almost no person
so very safe
can someone stop me from throwing my NAS out the window
*ima KILL IT!
I baught a Thecus N4100 from facebook, im trying to factory reset it due to forgetting my admin password
Ive had the system working before so i know its not physically broken
ive read the instructions on the website (the manual)
and a few threads of forums telling me different procedures
WTF do i do? its still locked up
The ip is locked with the ip of 192.168.43.52 and its suppose to be 1.100?
how are you trying to do that? DHCP?
ah
current im not connected to a network. my pc has an IP of 192.168.1.10? the nas has something completly different and i cant factory reset it to its default
currently*
seems like it didn't factory reset?
subnet mask of 255.255.255.0?
or /24 in CIDR format
.0
its windows default
i swear im missing something trivial here
can i ask a stupid question? do nas's have a maximum storage capacity that would stop it working?
you did the 30/30/30 thing?
no they shouldn't
unless it's some bad design
because i was looking at the Netgear RND4000
why I just run plain linux for a NAS
@clear igloo
AT&T lost another customer in the area
are they that crap?
im guessing your in america of overseas! in the UK we have like 20 providers
@minor girder connect your computer directly to the NAS
set a static IP on the same subnet as the NAS was configured on
that's what he did
probably a /16
he's trying to factory reset it
can't he just change IP settings?
well even houdini himself couldnt help us then
@tame carbon mate my issue is that i cant change any settings without it being factory reset? im not sure if there is a jumper on the inside or soemthing i could jump
you recon if i open it up i could just bridging circuits?
i cant seem to find an internal scematic
has the manual been consulted yet?
ok xD
they are translated English from Japanese lol
must be good
sometimes bsd
hmm so they have a cmos or bios
most of those embedded systems use something like uboot
hmmmm
usually a serial port on the board
but thats only to load an operating system onto the board
not to change passwords
all i have is two ethernets and a power
yeah thats on the inside
ah
sometimes theres not even a connector
you have to solder some wires onto the board xD
raspberry pi would be faster
whoah mate going way to fast for me
the pi 4b+ has 4GB/s memory bandwidth.
that's more than you'll be pulling out of that DDR2 controller
Guys, DD-WRT was giving me so much trouble last few updates I decided to go back to stock, this is a TP-Link Archer C7 V2, on stock firmware I can't seem to set the router IP to 192.168.10.10, it says I'm trying to put LAN in the same subnet range as WAN.
do you have two routers behind eachother?
are they on different networks?
PFSense to TP-Link
yeah you probably will want to not use the WAN port on the tplink
and disable WAN entirely
dhcp should be controlled by main router am i wrong?
I'd assume so yeah
I have DHCP on PFSense and disabled it on the TP-Link
@south blade yeah dont plug anything into the wan port on the tplink
your using one as a switch me thinks
set its network configuration on WAN to static and leave it unconfigured.
cant dhcp be set to auto?
DHCP is for the entire layer 2 network
@minor girder when a client connects to an ethernet network, by default it will send a UDP packet to 255.255.255.255
the dhcp server on the local network will then respond
iit doesnt use a range ?
the client then asks for an IP, and the server allocates one
the server can be configured to do whatever you want
you can for exmaple do mac binding
where the dhcp server assigns a lease to a specific mac
ok let me hackintosh lol
I use that heavily here
anyway crystal you were asking about nv2 and nstreme, there is not much point to them anymore
@tender hazel I looked up how to do the ptp stuff
*grabs popcorn
and there's like, various modes and protocols
*and poppron
there's really no reason to use anything but regular 802.11
even for ptmp?
if you have fixed ptmp with mikrotik CPE devices then there is a point to NV2
but it won't work with non mikrotik CPE devices (phones, laptops, etc)
that's the point
its between two mikrotik devices
and according to the doc
you get superior latency on those protocols
I was looking at NStreme specifically
Yeah, I think I had it set up as a switch basically when I had DD-WRT installed
Not sure how to do it on stock cause I had WAN set as a LAN I think before on DD-WRT, so I just disabled it's configuration type option, so no Static, Dynamic, or any of that. I just reset it cause I locked myself out somehow.
Dynamic is.you best option
I had it set to 192.168.10.10 before, and IPs started at 192.168.10.100 from PFSense DHCP
you might get a bit more out of nstreme, but be sure to test all three
Ugh I hate this network wtf
and it's not only latency you'll want to look at, but the throughput
@south blade honestly the amount of times one number has fkd me over lol
This guy setup clients with static IPs on the CLIENT side
best is DHCP reservation
And apparently the stupid DHCP server isn't pinging IPs before giving the lease
So this stupid client stole an IP from the ricoh
lol
@south blade what's your ip range?
I might be wrong on this assumption
Could it out of the range of ip addresses of the main router
If that's the one doing the dhcp
If it is out of the dhcp pool that's not an issue
it's only an issue if out of the subnet
Ok so his 255.255 us wrong
Its always good practice to try a different ethernet port or cable ;)
I would think the TP-Link firmware would get address by DHCP
DHCP reservation would be the best thing then
I had PFSense doing the DHC handing, trying to figure out how to get it to do it again.
I mean if you have a good link then I doubt it would be the cable
are we using two differnt setup menus here?
I don't know how to do it pfsense, but I know for sure there is no map static ip option like in EdgeMax
I think you need to copy and paste the MAC or something like that
@tame carbon they are basically discontinuing NV2 and nstreme
@tender hazel could you just hide the network by not transmitting the ssid?
I've seen threads like this before: https://forum.mikrotik.com/viewtopic.php?t=118129
shoot, I thought it'd be easy 5 mins to switch from DD-WRT to stock but run the same, LOL. I think I'm going to have to wait till late at night so no one will complain about me messing something up. Locked myself out disabling DHCP, setting the router to 192.168.10.10, even though I set gateway to PFSense 192.168.10.1
@tender hazel so less fault tolerant
the issue is that they never really improved nv2 for ac
same with nstreme
they are the same nv2 and nstreme they had before with wireless n on 2.4ghz
@tender hazel but can you really reach those rated distances from the brochure, with a 90 degree antenna, 1km ?
so when you use them you aren't taking advantage of some of the improvements between wireless n and ac
Its just new to me that 802.11 is used on such long distances
but the problem is you need a fixed cpe device on the other end to get that distance
it won't work with phones and laptops because their antenna is too weak
I know
@tender hazel I'm still working out a solution to that setup I might have to do next month
and we basically just want to buy a pile of those dual band mANTBoxes
@tender hazel those things have two chains, can you use them independently?
like, use only a single one of the 5GHz chains for the ptp
no
the problem is the interfaces show up based on the wireless chips that are in the unit, each chip = 1 interface
you can use both AP mode and ptp
what I would do is just take the 5ghz device out of capsman on those devices that you need to link
yeah but the place where that antenna most likely will be located
is also a dense area
use the 5ghz as a backhaul for those ones only, the rest you can use 5ghz as another capsman device
oh ok
so I have to have both 2.4 and 5GHz
then you may need a separate PTP link for that
could I just get a pair of those SXT's ?
yes
put them on a diff channel
yup
its like 150 meter at the most
60ghz would be better
what happens if there's leaves inbetween ? like trees
@tender hazel yeah but I dont want to get permit
do you really have to get licenses for 60ghz in germany?
that takes months.
yes.
paper pushers
@tender hazel ISPs reserve the right to use these
mere mortals cannot
@tender hazel but is 5GHz not fine in this situation?
it is but 60ghz is nice because no interference
what hardware would you recommend for such a link?
@tender hazel the thing is, would be nice xD
because the area that its going to
that is where all the youth gathers in the evening
and hangs out when parents are doing drinking and stuff
is there a document online somewhere that explains these 60ghz rules in germany? I can't seem to find anything about that
Richtfunk
you need to file this paperwork
oh
link got yeeted by the bot
There's links to pdf forms
you need to fill out
and 60GHz is not listed
I think 60ghz is not listed because it is unlicensed
wait hold on
@tender hazel check this ^
This is open to the public
57,1 – 57,8
58,6 – 58,9
GHz
and the law is only until 2023
so might be amended
point to multipoint you have to get a permit for anyhow
I don't know what that says
but that is obviously explaining the regulations around the 60ghz frequency
wow.
that website has Really old legislation
wtf germany
that was a 2013 law
and the one you just linked is 2020
@tender hazel holy fuk
316W ?!
are you able to use it unlicensed? what does it say regarding that?
and yes 60ghz needs to be really high power because in that particular band, atmospheric attenuation is quite high
the signal is extremely directional so that aiming is important, and power needs to be very high to cut through the atmospheric attenuation
but it means that you could have two separate 60ghz links running on the same frequency with the radios close by each other and not have to worry about interference
and what about people interference?
That stuff (the high freq transmitters) makes me physically sick
people interference?
as I said it is extremely directional, and the way 60ghz works, even a piece of paper or cloth would be enough to block it completely
yes, but it uses beamforming
its only 150m
the two dishes basically have hundreds of little antenna cells
yes, but with 60ghz you can get 1Gbps throughput for the same price
and use one of them for the ptp
or
I hook up one of those SXT's
@tender hazel 500mbit's is enough
with ac 80Mhz ?
I thought you said these things could do long distance :P
does the speed drop off that quickly
well you can get a bit more, but you aren't going to get 500Mbps
you can maybe get around 200 or 250Mbps
Im trying to use the calculator
you don't have to be afraid of 60ghz though, it is not dangerous to people
it is much better to use that for links that you can use that for b/c then you free up 5ghz spectrum for what you need it for
for instance with terragraph
terragraph is deploying 60ghz mesh on lampposts, like this: https://terragraph.com/
it is not high above people
@tender hazel https://mikrotik.com/product/RBSXTG-5HPnD-SAr2
can't I just test it out then?
i mean these things aren't that expensive
only supports wifi 4...
this one
yes you can test it out, but the 60ghz is so cheap, it is better
for this application
you are getting a worse product for no reason
well
the mikrotik 60ghz only draws 9 watts power
so I can't see if the radio is only drawing 9 watts how it would generate 316W signal
yes
this is the most inexpensive one
you can buy a pair of them
so which direction does it fire?
ooh
sick.
@tender hazel and it does regular 5GHz too?
they don't show the antenna patterns but it is highly directional
yes it does both
in one radio
oh, so switchable?
if there is some kind of physical blockage it fails over to 5ghz
or can you use both?
what can happen with 60ghz is in situations like torrential rain, if the link is longer, the rain can take the link down
do these run RouterOS?
the 5ghz is there is a backup in case there is rain
yes
or we once had a wireless link go down because someone put a billboard up in between the two radios
should I make sure I reserve the right frequency for the backup?
yes
the backup won't need a lot of bandwidth though because it is really just to ensure that at least something keeps operating
ye
"oh ye signal is slow"
"antenna is blocked"
is better than: wifi not working at all
yes
and yes they run routeros, and when you buy them in a pair, they are already preconfigured in a PTP link out of the box
can you have multiple?
do they automatically aquire this?
you choose the frequency, I think they picked some frequency as a default in the out of the box config
the lower 60ghz frequencies suffer the most from atmospheric attenuation, as you go a bit higher they will go longer distances
and this would skip 2x 80meters of cable
then between the poles where there's lots of forested areas
we can just dig cables
power is everywhere
so that's no issue
@tender hazel I might get one of those outdoor base boxes
and these can cover 80 meters no problem
err fiber boxes or whatever they are called
and run 3 long ethernet cables down the length of the camping
each cable goes to 1 pole with two sectors
you can do 60ghz over hundreds of meters
@tender hazel this is across a lake thats 30 meters wide
from an elevated position
there's clear sight
lakes can do weird things with wireless signals, just FYI
@tender hazel this is what the current ISP has hanging
you can try it
30 meters is not very long anyway
probably running over that short of a distance of lake will be fine
I can't tell what radios those are
I think they are metal 52's
@tender hazel its basically one entire ptp network
town -> hill -> this pole in picture -> house -> router -> more poles -> public wifi
this gear will be gone soon
ok
but they probably won't remove the pole
so you could stick equipment on that pole if it is helpful
so yes, the wireless wire cube is probably a great solution for that
only problem
this is probablyt where I'd put the sector antenna
this is close the pool area
so I might put two of those cubes on that pole
one from the house
yeah it should be fine, they aren't very big
and the other to the field area 80m away
I'd put a sector antenna ontop for local coverage
and then I need some kind of weatherproof device
to get more ethernet ports
I can't just daisy chain them all
xD
the mANT has SFP and ethernet
I assume I can use both
yes you should be able to use both at the same time
the only time you can't use both is when they are "combo" ports
but I think only the CCR routers have such "combo" ports
with SFP+ and 10GbE ?
@tender hazel that makes most of this quite easy
because there's a pavement inbetween the lake and the house
and I didnt want ot have to go underneath
that would be a whole week ordeal
so overtop is great
and once we are in the forested area, we can just lay a rugged cable
yup
so I would have three of them basically
so you aren't really losing any speed vs a cable, except in crazy weather conditions that would only happen a couple times a year at most
3 pairs
lol crazy weather conditions?
@tender hazel you havent seen bad weather there
couple years ago
they had a 1 meter floodwave of mud and water
a caravan ended up in the lake
along with a couple tonns of silt
and thousands of dead fish xD
the smell
wow
like, suddenly, the sky burst open and in a matter of minutes, dumped enourmous quantities of water
and the nearby stream of water that is usually calm
turned into a river
its a valley
so its hit or miss
we have heavy torrential rain suddenly in our area
if the cloud goes into the valley it compresses
cool
so we solved another problem then, and made this much easier for you 🙂
Yeah definitely
Only thing now, on one end, I might need to have a sort of outdoor switch
This is for the other direction
That's a bit overkill
16 ports is probably a bit overkill yes
Perfect
Yeah that's basically it
I need 3 cables to run to poles
And that 60ghz cube
With that box
the powerbox pro gives 24V output so you power the radios with that instead of injectors
the powerbox pro is better than the regular powerbox because it has a faster CPU so you can do software bridging if you need to - you wouldn't want to have to do software bridging with the regular powerbox
so I would spend the extra money on the powerbox pro vs. the powerbox, even though they are otherwise similar
Oh yeah more flexibility is nice
the powerbox pro has an SFP port too, and the powerbox does not
Cool, 3x ptp two power boxes and a 8 or so sector antennas, id say less than 2500 in total
Not bad all things considered
I was just gonna use an RB4011 as controller
it will be a fair bit of work configuring everything, but it should be really nice once it is done
you'll have wireless that will be the envy of other campsites
xD
Currently guests complain in online reviews that the wifi is a ripoff
But we dont sell those vouchers, we're at the mercy of the isp
a ripoff? do they have to pay for it? oh goodness
Yes
Even the private net isnt flatrate
Its crazy
But its either that, or go bust
how were you going to handle it with the new system - just give free?

