Tagged VLAN - losing connection until ARP is cleared

[TheLilLebowski]

Hi all,

Attempting to set up a tagged VLAN on a 2615-8 switch.  I will be traveling out of the country to install this, so I'm doing some testing now, before my trip, to eliminate any potential issues while I'm on-site and my time is limited.

The default VLAN 1 is untagged on all ports, IP address is set to 192.168.201.1.  VLAN 251 is tagged on all ports, IP address for switch on VLAN 2 is 192.168.251.1.  PC #1 is connected directly to the switch and has a static IP address on VLAN 1 (192.168.201.23).  PC #2 is connected directly to the switch, has a static IP address on VLAN 2 (192.168.251.23) and has the VLAN 251 tag assigned within the NIC options.

Everything seems to be working fine initially - I am able to ping all four addresses from any one of the devices.  The problem comes in when I change ports.  If I plug either of my PCs into a different port and immediately try a ping, I am no longer able to ping from one PC to the other.  If I wait about a minute and try to ping again, my first ping packet is successful, but all other subsequent packets fail.  Again, if I wait about a minute, the same thing happens - first packet is fine, everything else fails.  I am still able to ping both PCs from the switch, using either VLAN as the source.  If I move back to the original port, pings are again successful.

Thinking this is an ARP issue, I clear ARP cache and everything works again.  However, if I just wait for the ARP cache timeout (20 minutes), the problem is still there - I actually waited about 40 minutes with no change.  Tried changing the ARP timeout to 1 minute also.  The odd thing with this is that if I look at the ARP table immediately after switching ports, the ARP table is updating correctly.  

I'm sure this is related to the tagged VLAN - if I put both PCs on VLAN 1 and switch ports around, there is no issue.  I'm just not sure what is going on.  The fact that one ping packet works and then all others fail boggles my mind.  No matter if I wait a minute, 5 minutes or 40 minutes, it happens every time until I manually clear ARP cache or reboot.

Is there something I'm missing with the tagged VLAN?  I've done this before on other HP switches without these issues...  Thanks in advance for your help!

Latest firmware installed, config is attached.

[Vince-Whirlwind]

I don't think you have proven that the tagging makes the difference:

if you put them both on VLAN1, then they are communicating directly using each other's MAC address.

if you have them on different VLANs then they do not address each other's MAC address.

So the issue occurs when the *switch* is trying to find the destination IP address, and NOT when the PCs are trying to find the destination MAC address.

A clear ARP refreshes the L3 table in the switch.

Switches these days perform routing in hardware, not CPU, so the issue would seem to be in the L3 switching.
It therefore seems very unlikely that you could do anything wrong in config to cause this.

[TheLilLebowski]

Thanks for the reply.  I did try putting them on separate VLANs without tagging - ports 1 - 4 untagged on VLAN 1, ports 5 - 8 untagged on VLAN 251.  Everything works as expected in this configruration.

I also tried uplinking another switch (2510G-24) to see if this changed behavior at all.  2510G is set up with the same VLANs.  All ports on both switches are untagged on VLAN 1, tagged on VLAN 251.  I can move ports around the 2510G without a problem.  If I move ports on the 2615, the problem comes back.  Doesn't matter which PC is plugged into which switch, the result is the same.

Looks like a call to HP support is next.  I thought I'd throw it out here first to see if there was something that I was just missing in my config.

[Vince-Whirlwind]

Just out of interest - is the faulty behaviour any different (on the 2615) if the device you repatch is rebooted?