5

I have a star topology consisting of 12 switches. Recently, we began running low on available IP addresses (192.168.100.0/24) and I was asked to open up a new scope (192.168.50.0/24) on the DHCP server.

The scope has been created, I tied it to the same VLAN 50 on all switches, tagged and untagged the appropriate ports and everything seemed to be fine. The clients correctly grab their IP address from the newly designated scope, and they have full network access.

The problem we are seeing is that when a host tries to communicate with a host on the other VLAN, there is anywhere between 0-10% packet loss. If I run a constant ping across the VLANs and it isn't dropping any packets, I will get 3-4ms per attempt and then a 30-40ms packet every 15-20. Any host that communicates with its own VLAN is flawless.

The switches are mostly HP Procurve 1920-24G, with a couple of 1910-24G models that haven't been replaced yet.

I have checked on multiple occasions to make sure that there isn't a loop in the network, because the switch logs all seem to show ports intermittently changing their status to DOWN and back to UP right away over a period of a few seconds. Sometimes this will happen every 90 minutes or so, and the port is not always the same.

It just seems quite strange that there are only issues when the communication crosses over to the other VLAN.

I can upload any configs or output, if needed.

  • Instead of two scopes and VLANs, could you use a wider scope? For example 192.168.100.0/22, gives you a subnet of 1024 addresses, from 192.168.100.0 to 192.168.103.255 – jcbermu Oct 29 '15 at 8:51
  • The network is currently setup to give 192.168.101.*/24 addresses to VPN users. I was hoping to not have to touch this, but I might not have a choice. It won't be hard to move that scope, so this is may be the route I go. – MooseBalm Oct 29 '15 at 13:27
  • @jcbermu - I ended up just doing this. We ran out of time to find a solution and needed an immediate fix. It ended up being more work to make it happen, but it removed any inter-VLAN routing, and therefore the problem. – MooseBalm Oct 30 '15 at 20:44
  • Did any answer help you? If so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you can provide your own answer and accept it. – Ron Maupin Aug 7 '17 at 0:40
3

If the problem only happens between VLANs, you need to look at the router which is used to route between the VLANs. It may be overloaded on its CPU or buffers, but that is the common point between VLANs, and that is the place to start your investigation.

  • I can at least verify that it is not the CPU. It hasn't exceeded 25%. If I can find a way to check the buffers on this model, I'll report back with more info. – MooseBalm Oct 28 '15 at 15:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.