Moving from PVST+ to Rapid PVST+

[hbaytie01]

Hi,

I faced an incident with one of our customers where the AVAYA phones were not receiving DHCP IP on the access switch, only this avaya voice vlan wasn't working and all the other vlans were receving DHCP. The access switch was connected to the Core through Ether-channel (PAGP), the only thing I thought about is to use Active/Passive (LACP) instead and test it. However, once we removed the ether-channel on the access switch the whole network went into a loop and nothing became accessible anymore until we rebooted the access switch. After checking the Spanning Tree, i noticed that the core has Rapid-PVST+ with this configuration:

spanning-tree mode rapid-pvst
spanning-tree extend system-id
spanning-tree vlan 4,7-8,10-26,30,41,59,90,92,100-102,105,111 priority 24576
spanning-tree vlan 112,198-200,205,223,250-255 priority 24576

And for the access switches, they are left as default PVST+. 

I'm thinking to fix the Spanning tree for the network to make all switches use rapid-pvst and force the core to be the root bridge by using the commands:

Core(config)#spanning-tree mode rapid-pvst 

Core(config)#spanning-tree vlan 1-4096 root primary 

and for the access switches only using the command: spanning-tree mode rapid-pvst 

My questions are:

1- Once i apply the above commands to the core will all the network be down and for how much time?

2- Once i apply the command to the access switch to move them to rapid-pvst will be they become down and for how much time?

3- Will fixing the spanning tree solve the DHCP problem i faced with avaya voice vlan and will that prevent the loop that happened once i removed the ether-channel on one of the access switches?

[Deepak Kumar]

Hi, 

Before going to your answer, I want to ask few questions:

1. Are the LLDP protocol enabled on Switches?

2. Where is DHCP configured for Avaya Phones?

3. Are the IP Office Manager reachable from Voice VLAN?

4. If DHCP is configured on another device as Core switch or Server, did you configured required DHCP option for Avaya? Did you assign IP helper-address (If required)?

Now I am coming to your answer. 

1- Once i apply the above commands to the core will all the network be down and for how much time?

 Ans: It may be downtime (Flap) approx the 15 seconds to 2 minutes, the Same time it depends on STP timing.

Recommended: After office hrs or take downtime approval.

2- Once i apply the command to the access switch to move them to rapid-pvst will be they become down and for how much time?

Ans: It may be downtime (Flap) approx the 15 seconds to 2 minutes, the Same time it depends on STP timing.

Recommended: After office hrs or take downtime approval.

3- Will fixing the spanning tree solve the DHCP problem i faced with avaya voice vlan and will that prevent the loop that happened once i removed the ether-channel on one of the access switches?

Ans: It is not guaranteed to resolve your issue. Yes STP will do it own task, same time keep in mind that if your VLAN configuration is ok then EtherChannel will not make any issue in DHCP process (if other security features not enabled such as ARP, DHCP snooping etc.). 

[hbaytie01]

Hi Deepak,

Thanks for your reply, 

Regarding the LLDP, there is no configuration done for it, so i think by default it is disabled.

The phones can ping their manager and the thing is that if you remove the ethrchannel which has trunk mode and connect one  uplink as a trunk also to the same switch, the phones will receive dhcp. They have windows dhcp sever and there are ip-helpers for vlans.

There are also no security features enabled.

[Deepak Kumar]

Hi, 

First of all, Avaya needs LLDP enable on the switch to identify the Voice VLAN and other parameters. 

If you are sure (Checked in your network) that Phones are getting IP after removing the EtherChannel then you have to check EtherChannel configuration. I am sure there is some misconfiguration on EtherChannel interface.

Regards,

Deepak Kumar 

[hbaytie01]

Regarding the LLDP, I don't think it is causing a problem since once we remove the two cables(Ether-Channel) and connect one cable(Trunk), everything will work normally. 

For the EtherChannel configuration, it is a normal trunk same as the rest of floors. This is confusing because it's not logical.

[Deepak Kumar]

Hi, 

Is it possible to share ethernet status and configuration from both switches?

Regards,

Deepak Kumar

[Joseph W. Doherty]

Although I would recommend you move ahead having all your devices use rapid-PVST, rapid and non-rapid should interoperate, so not having rapid on all your L2 devices now may not have been the issue (or cure for question #3).

Regarding you questions 1 and 2, it will probably impact your network from a few seconds to up a couple of minutes. As rapid-pvst takes over, L2 convergence will take less time (i.e. seconds - which is why it's named rapid).

BTW, Deepak notes you need LLDP to use Avaya phones. From personal experience, I've used Avaya phones on Cisco switches that didn't support LLDP. (However, I have seen old firmware on Avaya phones and on Cisco switches, together, to preclude Avaya phones from booting.)

As to why your Avaya phone don't grap DHCP, hard to say, as the phones can be programmed on the phone itself or via initial DHCP option strings. Ether set incorrectly could cause issues.