Jump to content

Recommended Posts

Posted

This one is doing my head in now.

 

First reported Monday, devices connect with no internet. Fairly isolated - maybe 2 people, and the problem went away when they went somewhere else in the building. Yesterday, alll the HLTA and Intervention staff laptops wouldn't connect.

 

Symptom: Windows devices get a self-assigned 169 address which cannot be cleared or reset. Android/Chromebooks refuse to connect - DHCP lookup failed.

 

Once the windows device has a 169, even plugging a cable in results in a 169 on the ethernet as well until it gets a valid IP on it's wireless interface.

 

I've got about 80 devices that are connected and not causing issues, but there is nothing obviously different about them.

 

The really weird thing is that if you go somewhere else in the building to another AP the device then usually connects like there has never been an issue.

 

Ideas?

Posted

Not directly - we don't use vlan 1, we are using vlan 2 for everything computing on this site and all switch ports are set native to vlan 2 unless they serve a printer or a door controller.

  • Like 1
Posted

Have you tried wireshark on one of the windows devices? You should be able to use the bootp or dhcp filter to show only those packets.

 

I had a weird issue on my windows laptop yesterday where the laptop didn't even seem to send the DHCP request on the cabled connection, but was fine on the wifi.

 

We use Unifi for switching, but not wifi.

 

If the other 80 devices already have a dhcp lease, then it might just keep using that until it runs out.

 

The other thing to check is the dhcp settings on each network on the unifi controller and/or router, things like DHCP forwarding, DHCP Guarding, DHCP Snooping (Rogue DHCP server detection) etc

  • Thanks 1
Posted
53 minutes ago, Chris_Cook said:

Have you tried wireshark on one of the windows devices? You should be able to use the bootp or dhcp filter to show only those packets.

 

I had a weird issue on my windows laptop yesterday where the laptop didn't even seem to send the DHCP request on the cabled connection, but was fine on the wifi.

 

We use Unifi for switching, but not wifi.

 

If the other 80 devices already have a dhcp lease, then it might just keep using that until it runs out.

 

The other thing to check is the dhcp settings on each network on the unifi controller and/or router, things like DHCP forwarding, DHCP Guarding, DHCP Snooping (Rogue DHCP server detection) etc

Not got as far as Wireshark yet...

 

I've tried a ipconfig /release and /renew on a device that is working and it just picked up another IP in seconds. Another non-working device, literally sat next to it connected to the same AP complains that that the DHCP has timed out.

 

If I pick the device up and walk to the other school across the road, which is the same wireless network, it picks up an IP like there is no problem.

Posted

What model ap and firmware?

I only have Unifi at home but I have seen weird issues occasionally (UDMSE does DHCP) and I ended up trying a lot of things and it eventually worked. You could see if there are any logs in the AP itself?

Posted

FWIW, and not Unifi,  we've been seeing increasing numbers of random wifi problems over the last month, it seems likely there have been a windows / driver update that may be complicit.  (mostly Atheros chipsets). Nothing of a scale to cause wailing or nashing of teeth, but definitely something we're keeping an eye on.

 

Part of our problem is/was some of the Atheros chipset drivers have "randomise mac" flag (which overrides the Windows setting), and this exhausted a DHCP scope recently.

 

 

  • Thanks 1
Posted
56 minutes ago, ZeroHour said:

What model ap and firmware?

Mainly Nano HD but a couple of U6 Mesh Pro. 

 

6.7.17 on the Nano and 6.6.77 on the U6 Mesh. Both apparently up to date.

Posted

Uggghhh, right, apologies everyone who replied to this topic & to Unifi, not your fault, not this time!

 

Turns out the DHCP server was exhausted.

 

What 'appened woz...

 

in the Easter holidays and under instruction from our Trust I dialled down the DHCP scope on the Smoothwall from an admittedly excessive /16 to a /24 with an 8hr lease time and 250 adresses available via DHCP. This particular site has around 150 normally connected devices on this with an ocasional usage of about 40 additional devices.  At the weekend the lease time changed back to 2 days. I'm still not clear if this was a deliberate change, an update, or an automatic settings overwrite from the parent node.

 

A combination of factors, including a large Trust meeting where around 30 people used the Trust's network which sits on our scope and the use of a suite of Chromebooks that don't get used often, caused a higher than normal number of clients to connect Monday into Tuesday and hang onto IPs, thus by Wednesday lunch time the scope was operating one-out-one-in. 🙄 Late Wednesday night (literally at about 8pm) the Trust finally got back to me and we agreed that 2 days was excessive, dialled the lease time back to 8 hours and everything is fine this morning, around 50 free leases. We've also agreed that the Trust will put their network on it's own scope & VLAN.

 

 

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...