Forum Discussion

Guide

Feb 01, 2019

Solved

Voice VLAN command - LLDP seems to stop working

I have a strange issue with the voice vlan command that I'm trying to isolate between our switches and Shoretel phones, but it seems there's no debug options for LLDP on this switch. We use the...

Troubleshooting

tubesfarmer
Oct 06, 2019
FWIW this issue seems to be fixed. Originally the config was

interface 1/0/1
voice vlan 10
switchport mode access
switchport access vlan 20

But the switch would randomly stop passing the tagged VLAN traffic. I changed the config to:

interface 1/0/1
voice vlan 10
switchport mode trunk
switchport trunk native vlan 20
switchport trunk allowed vlan 10,20

I haven't had a single DHCP failure/VLAN tag drop since moving to this config.

This behavior seems to only occur on the M4300 series - the M4100 actually rejects switchport mode access as a valid option when using voice VLAN (I have those access ports configured in general mode but I'll be updating them later for the sake of consistency). Maybe it would be nice in future updates to prompt users to use a trunk config when enabling voice VLAN on switchports so they don't spend their time flailing about like I did (this problem lasted for like, seven or eight months!). :)

To be clear, it wasn't LLDP failing to maintain the voice VLAN tag, it was the switch refusing to pass tagged traffic and the phone dropping the VLAN tag after DHCP repeatedly failed (which is normal Mitel/Shoretel behavior).

If anyone else googles this problem hopefully they find this post. :) :womanvery-happy:

DaneA

NETGEAR Employee Retired

Feb 04, 2019

Hi technotechnotec,

Were you able to update the firmware of the M4300-52G-PoE+ switch to v12.0.7.10? If yes, what are your observations?

Let me share the link and it might help:

Auto-VoIP and LLDP-MED Guide

Regards,

DaneA

NETGEAR Community Team

DaneA

NETGEAR Employee Retired

Mar 03, 2019

Hi technotechnotec,

Thank you for the update. It would be best that you open a chat or online support ticket with NETGEAR Support at anytime. Be sure to attach the tech support file from your M4300-52G-PoE+ switch and indicate what port number the the issue has occurred and this might be possibly escalated to our engineering team for investigation.

Regards,

DaneA

NETGEAR Community Team

IT-Adm1n
Aspirant
Mar 04, 2019
Just wanted to report that we are experiencing the same exact issue, but with two independant GS752TPv2 switches located in two different office sections, interconnected by a single up-link, both with the latest available firmware v6.0.0.45.

We have a combinasion of mostly Yealink VoIP phones and a few GrandStream units.

Everything worked fine for a few day, but the problem reoccurs after less than a week.

I can confirm the observations that the OUI based Auto-VoIP still sets the VLAN, but *untagged*, which breaks the configuration until said port is temporarilly set to untagged. Temporarilly untagged meaning until the next switch reboot that is as, when it works, LLPD indeed seems to consistently properly configure the phones to set and activate their VLAN tagging.

Rebooting a switch definitely cannot be done during business hours so this quite a hassle, especially the time required to insure that all phones indeed came back onoine properly afterwards.

One of the switch has fewer devices connected to it so I could troubleshoot during normal hours after confirming everyone were out to lunch and right after the switch reboot it started to properly communicate the LLPD configuration changes:

*Mar 04 2019 09:16:48: VLAN-6-VOICE_MBR_LLDP_ADD: Interface GigabitEthernet23 added to voice VLAN 5 by LLDP
*Mar 04 2019 09:16:48: VLAN-6-VOICE_MBR_LLDP_DEL: Interface GigabitEthernet23 removed from voice VLAN 5 by LLDP

When the issue is present all of those LLDP events stop happening, only the link up/down events and related blocking/forwarding entries appear.

I would be fine with just issuing a restart command to the LLDP service/daemon, but I understand that this level of access may not be possible with the lower class model we have.

Hopefully there is a fix in the works and/or soon(ish).

If I need to open a support case I will, but no need if this is now a known problem etc.
technotechnotec
Guide
Mar 28, 2019
That's essentially what I've had to do. This issue is completely random and unfixable without a reboot. Oh well, here's hoping that engineering "may" help. :)