Pings timing out under strange circumstances - to DC's from only certain PC's

[mikes]

Hi guys not sure of the right place to put this, we have been getting odd issues with our network - I think it started when we had new switches fitted by our LEA but can't be sure.

I get the odd issue with staff complaining about network dropouts; and the symptom I get is this:

 

My workstation will lose connectivity to certain servers eg. DomainC-02 - I'll lose my RDP connection and pings will time out for 3 ping attempts. But I can ping other servers fine and connectivity stays up.

However the odd thing is that other workstations around the site, can continue to ping the affected servers perfectly fine - they won't lose any packets. Also the time I will lose ping connectivity (to say DomainC-2 and Domain-C1) is usually the same time but one can be a few seconds earlier than the other, or one might stay down 5 seconds longer than the other etc.

 

Also these servers are virtual servers, and I will lose ping traffic to say 2 of the servers on that vhost while 5 other servers on the same vhost are fine so it's not a vhost issue?

 

When the ping from my workstation drops out to the affected servers, I can have a ping running from the servers back to my PC - the pings outbound to my PC go through but the Milisecond response time shoots up to ~200ms instead of 1ms for the period of 5 - 15 seconds or so that this happens.

 

Our switches are run by our local authority so it limits the diagnosing I can do, however they say they see no problems with the switches so I am stuck there. (We also have issues with non-workstations e.g. voip phones and printers taking 5+ minutes to obtain DHCP leases, and our voip phone setup often phone calls will be silent for the first 7 seconds but these may be unrelated)

 

We have Aruba switches, anybody know what might be causing this?? It's a hard thing to even search for describing it can make me go round in circles

[psydii]

Sounds like a conflict between LACP and VMware IPHash Load balancing.

[FN-GM]

Do you loose pings to something other than your virtual servers (maybe you have wireless controller, firewall, router or similar)? In addition are you loosing pings to your physical hosts? Do your virtual machines migrate between physical hosts or are they static? Edited March 20, 2023 by FN-GM

[mbedford]

Your symptoms are very similar to what I would expect to be seeing with a network loop or broadcast storm - https://en.wikipedia.org/wiki/Switching_loop

 

However, there are protections in place on modern switches and routers to mitigate the effect of this problem, assuming they are turned on in the device configuration (BPDU, Loop Guard etc).

 

It's a time consuming task, but I would physically check that there isn't a switch plugged into itself (or a device port in the classroom\office plugged into another port) creating a loop back to the switch.

 

Given that the LEA have said they cant see anything untoward though, it probably isn't this.

[mavhc]

Agreed, check spanning tree is on first. Also log into switches to test pings, see if you can narrow it down to certain switches.

 

Is there any limit on mac addresses for the ports the VM server is plugged into?