[Closed] System and network slow, DHCP problem??

Mellowlynx

I can never understand why people just "try things" to see if it works. Find the fault and fix it.

I'm not just trying, I have the same type of network at home and that runs fine.
and here we do need a guest wifi, we do not wand the payment terminal and POS system on the same network as the guests.
But yes, I'm not a pro… I do this for "fun" but this is getting annoying.

The problems started after a power outish.
At the power restore the old cases power supply and mainboard broke.

Derelict

Do you have the same SSIDs on the two wifi APs?

(I also don't understand why someone would hire someone who does this for "fun" when the problem is as important as PoS segmentation from guest networks, but that's probably just me.)

Mellowlynx

@Derelict:

Do you have the same SSIDs on the two wifi APs?

Nope both have differed names and the private one has a WPA2 pass
Guest is open.
But the owners do switch between booth.

@Derelict:

(I also don't understand why someone would hire someone who does this for "fun" when the problem is as important as PoS segmentation from guest networks, but that's probably just me.)

They don't hire me, that ask of i know a solution for there request.
as there friend I showed them my setup and ask me to get it installed.

Everything was running fine for almost two years. until the blackout.

Mellowlynx

Back on topic, re2 is now both wifi networks and for now everything seems to be working.

let's hope it will stay this way.

Derelict

Umm. So your private and guest networks are now on the same subnet? Thought you wanted them isolated.

Since you refuse to provide requested details and just want to plug things into different ports instead of actually design it in a correct, secure manner, I guess a "good luck" is in order.

Mellowlynx

@Derelict:

Umm. So your private and guest networks are now on the same subnet? Thought you wanted them isolated.

Only the wifi, lan is still a separate on and on there is the POS and Payment gateway.
And it should be temporarily

@Derelict:

Since you refuse to provide requested details and just want to plug things into different ports instead of actually design it in a correct, secure manner, I guess a "good luck" is in order.

I may have missed something, but I do not see any data requests in this topic.
Else I would post them no problem, I want to fix this.

I think I'm misunderstood, I do not want to " just plug things into different ports" I do what to fix this, but the Lan is working fine so I can't just shutdown the network for the day. They are open all day and I'm trying to work without giving them problems.

Also I did check the lan today everything is fine, from the music pc to the pos everything is connected the right way.
So no, I have no idea why a phone would request a ip on both networks ate the same time.

Again, If I can provide you some information that can make the situwation or problem more clear,
let me know and I will post it.

Derelict

Yeah. Make a diagram.

If you have both wifi networks on the same subnet, one open and one with WPA, they might as well both be open. They're the same network.

Mellowlynx

Guess I missed that one, but here you go.

This how is was this morning!

Derelict

Back on topic, re2 is now both wifi networks and for now everything seems to be working.

So you moved your private wifi to re2? Now it's the same network as your guest wifi. Might as well not have a WPA passphrase on it.

And how did you do that without adding a switch to re2?

Mellowlynx

@Derelict:

Back on topic, re2 is now both wifi networks and for now everything seems to be working.

So you moved your private wifi to re2? Now it's the same network as your guest wifi. Might as well not have a WPA passphrase on it.

And how did you do that without adding a switch to re2?

Yes I did, but once again it's temporarily.

I did add one, the image is the situation before moving the wifi.
aka the situation the problems started.

Mellowlynx

BTW, I still see this


Oct 9 22:12:09	dhcpd: DHCPREQUEST for 192.168.2.18 from 20:16:d8:a6:15:8a (Nicole) via re2
Oct 9 22:12:09	dhcpd: DHCPACK on 192.168.2.18 to 20:16:d8:a6:15:8a (Nicole) via re2
Oct 9 22:12:09	dhcpd: DHCPREQUEST for 192.168.2.18 from 20:16:d8:a6:15:8a (Nicole) via re2
Oct 9 22:12:09	dhcpd: DHCPACK on 192.168.2.18 to 20:16:d8:a6:15:8a (Nicole) via re2

Should be just one request and one pack right?

Derelict

Don't know why the client is requesting twice. Either the client is requesting it twice or your network is sending it to the DHCP interface twice (like there's a layer 2 loop or something.)

Mellowlynx

@Derelict:

your network is sending it to the DHCP interface twice (like there's a layer 2 loop or something.)

The switches are both unmanaged, so is there a way this could be done with settings in pfsense??

divsys

Just a small comment from someone who's seen some "odd" network situations over the years.

It may be worth your time to check EVERY cable connection in and out of the router, wifi's , and devices to make sure they actually match the diagram you've drawn.

I've seen cases where in a small setup like this, an "extra" connection ends up causing a network loop. Especially if someone else tries to desperately "fix" a network problem while you're not around.

It shouldn't take too long and is well worth the effort to make sure there isn't a phantom connection causing grief.

Just my $.02

Derelict

@rcktboy:

@Derelict:

your network is sending it to the DHCP interface twice (like there's a layer 2 loop or something.)

The switches are both unmanaged, so is there a way this could be done with settings in pfsense??

Not if you didn't configure a bridge.

Good advice above. Look at everything connected to both switches. Do some packet captures on re1 and re2. If you see any traffic that's supposed to be on the other interface you know you have a problem.

Mellowlynx

@divsys:

It may be worth your time to check EVERY cable connection in and out of the router, wifi's , and devices to make sure they actually match the diagram you've drawn.

I did check all the cable connections, and they check all out.
Today I will check and monitor the WiFi and change the password too.

The networks looks more stable and not as match double DHCP requests anymore.

Mellowlynx

The problems are still there :(
And the wifi has a different password and only the necessary stuff is connected right now.

Still the day before yesterday when it got busy in the cafe and more devices connected the network slowed down big time.

To rule out the system itself I put my old system there yesterday and hope this will give is more information.
The system has run here for more then 2 year without any problems so I hope it will there also.

But it this got thinking, and I checked also my DHCP logs,
And have to say that there is allot of stuff happening, even when most stuff is off and I'm not home.
I get that there is always network communication, bus also here allot of (in my eyes) unnecessary request.

Then I checked a other system,
And I see this


Oct 17 14:51:30	dhcpd: DHCPDISCOVER from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:51:30	dhcpd: DHCPOFFER on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:51:31	dhcpd: DHCPREQUEST for 172.16.32.125 (172.16.32.1) from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:51:31	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:51:40	dhcpd: DHCPREQUEST for 172.16.32.125 from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:51:40	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:52:39	dhcpd: DHCPREQUEST for 172.16.32.125 from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:52:39	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:55:25	dhcpd: DHCPREQUEST for 172.16.32.125 from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:55:25	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:56:24	dhcpd: DHCPDISCOVER from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:56:24	dhcpd: DHCPOFFER on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:56:25	dhcpd: DHCPREQUEST for 172.16.32.125 (172.16.32.1) from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:56:25	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:57:22	dhcpd: DHCPREQUEST for 172.16.32.125 from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:57:22	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:57:22	dhcpd: DHCPDISCOVER from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:57:22	dhcpd: DHCPOFFER on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:57:23	dhcpd: DHCPREQUEST for 172.16.32.125 (172.16.32.1) from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:57:23	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:58:25	dhcpd: DHCPREQUEST for 172.16.32.125 from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:58:25	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:58:25	dhcpd: DHCPDISCOVER from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:58:25	dhcpd: DHCPOFFER on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:58:26	dhcpd: DHCPREQUEST for 172.16.32.125 (172.16.32.1) from 8c:2d:aa:e0:90:2c (Susans-iPad) via re2
Oct 17 14:58:26	dhcpd: DHCPACK on 172.16.32.125 to 8c:2d:aa:e0:90:2c (Susans-iPad) via re2

The last system is a completely other situation and system and I think it's strange that it's happening there also.

Do others have there busy DHCP servers/logs also??
Or is there a maybe a bug??

Derelict

It is NOT a bug. You have something fundamentally wrong with your network.

I have a pfSense 2.1.5 system installed in a 2200-room hotel with 4000-ish simultaneous captive portal clients every weekend, all configured via DHCP.

And another serving about 500ksqft of meeting, exhibit, and banquet space. Multiple VLANs, multiple DHCP scopes, multiple SSIDs. It all just works.

When are you going to listen?

I don't see anything fundamentally wrong with those logs.

You'll have to ask Apple why the device is performing a DHCPDISCOVER/DHCPREQUEST so often, unless your lease time in the scope is 10 minutes.

My guess is you're trying to use cheap-ass wi-fi gear in a hospitality setting and it's falling on its face.

johnpoz

"Or is there a maybe a bug?"

A bug in your client maybe.. Cleary susans ipad likes to keep asking for IP every couple of minutes.. Getting it and answering it with ack, and then asking again and again and again…

Its quite possible she is bouncing up and down on the wifi so keeps wanting to renew or get an IP, etc.

here is my phone for example..
Oct 17 05:00:48 dhcpd: DHCPACK on 192.168.2.225 to ac:fd:ec:62:34:97 (Johns-Phone) via em2
Oct 17 05:00:48 dhcpd: DHCPREQUEST for 192.168.2.225 from ac:fd:ec:62:34:97 (Johns-Phone) via em2

Oct 16 17:00:25 dhcpd: DHCPACK on 192.168.2.225 to ac:fd:ec:62:34:97 (Johns-Phone) via em2
Oct 16 17:00:25 dhcpd: DHCPREQUEST for 192.168.2.225 (192.168.2.253) from ac:fd:ec:62:34:97 (Johns-Phone) via em2

So my lease is 86400 seconds or 24 hours.. So 5pm on 16th it renewed, and just as you would expect 12 hours later it renewed it again.. Not every couple of freaking minutes request, discover, etc. etc..

Derelict

Or the wi-fi coverage sucks so bad the client is constantly dropping in and out of association. Every time it rejoins it has no choice but to ask for DHCP again.

But it must be a bug in pfSense/ISC dhcpd.

ETA: (which is pretty much what you said in the part I didn't read the first time. :/)