First a little background. Our network is primarily Cisco based and as such we run VTP. Recently we bought the following:
Dell M1000e chassis
2 x M6348
2 x 8024-k
4 x M620 blades.
1 x 6224 switch
On the existing network all management is performed on VLAN 1 (10.44.1.1/21). Existing VLANS include our training VLAN 32 (10.44.32.1/21). No stacking is in place at this point, all switches are stand alone.
This is what I've done so far starting with a trunk from the existing cisco 6509 to the Dell 6224
Cisco 6509 - Link to Dell 6224
Configure 2 ports channel group 2
Protocol LACP
Configure port channel 2 switch mode trunk
Dell 6224
Configure 2 ports, channel group 1
(Link to cisco) Configure Port Channel 1, switchport general allowed vlan 32
Configure mgmt ip 10.44.1.3/21
Create VLAN 32
(Link 1 to M6348(1)) - Configure 2 ports, channel group 2
Configure Port Channel 2, switchport general allowed vlan 32
(Link 2 to M6348(2)) - Configure 2 ports, channel group 3
Configure Port Channel 3, switchport general allowed vlan 32
Test 1
Putty onto 6224 and ping existing infrastructure = no problems. Can ping across sites and networks.
Dell m6348's
Configure mgmt ip 10.44.1.4/21 (first switch) and 10.44.1.6/21 (second switch).
Create VLAN 32.
Create LAG using external ports 33/34 configured for channel-group 1 mode active
Create port channel 1, switchport general allowed vlan 32
Connect two cables from external ports to 6224 channel group 2
Connect two cables from external ports to 6224 channel group 3
Configure interface 1/0/1-4 as switchport general allowed vlan 32
Test 2
Putty onto 6224 to confirm LAG is working. Ping MGMT interface of M6348 no problems.
Putty onto M6348. Ping 6224 no problem. Ping existing infrastructure = no problems.
M620 Blade 1
Power up and configure ESXi management network for VLAN 32 IP address 10.44.32.150/21
M620 Blade 2
Power up and configure ESXi management network for VLAN 32 IP address 10.44.32.154/21
Test 3
Use iDrac to connect to console of blade 1
Using ESXi shell and ping blade 2 = fail
Using ESXi shell ping switch = fail
Using ESXi shell ping existing infrastructure = fail
Troubleshooting steps taken
From M6348 console ping blade 1 = Fail
From M6348 console ping existing infrastructure = Pass
From M6348 create interface vlan 32 and give it an IP address
From M6348 ping blade 1 = Pass
From M6348 ping existing infrastructure on VLAN 32 = Fail
From M6348 ping existing infrastructure on anything else = Pass
From Blade one console, ping switch = Fail
From Blade one console, ping blade 2 = Pass
So after that epic list, my questions are:
What am I doing wrong?
Why can't the blades get as far as the M6348 without the interface?
I've spent 3 days going over this and just can't see what the problem is.Next week I'm going to try and take the M6348 uplinks directly to the Cisco 6509, effectively removing the 6224 from the equation but that won't be as neat.
thanks for the response. I messed around quite a bit since posting this and eventually I managed to resolve the issue yesterday using the document linked above "Deployment of Dell M-Series blade switches in a Cisco Catalyst Network" <- This is a great read! A lot of my worries about adding power connect into our environment were removed because of this.
I started again with my configs and was much more methodical in working through the process.
Step 1: Create lag between 6224 and 6509, configure trunking and confirm working on something other than VLAN1
Step 2: Create lag between M6348's and 6224, configure as switchport general and tagg traffic
Step 3: Configure Each blade interface as trunks
It really was that simple. I got confused because from the 6224 I could ping everything on our network as long as it wasn't connected directly to the 6224. The reality was I hadn't been explicit enought with the port channel create between the Cisco and powerconnect.
You may think about stacking the two M6348 switches. Is the traffic coming from the blade tagged?
Do you have a PVID set on the general switchport connecting to the blade?
switchport general pvid
Use the switchport general pvid command in Interface Configuration mode to configure the Port VLAN ID (PVID) when the interface is in general mode. Use the switchport mode general command to set the VLAN membership mode of a port to "general." To configure the default value, use the no form of this command.
This would tag any traffic coming from the blades to the VLAN designated from the PVID configured on the port.
ollyu4
2 Posts
1768
1
Posted October 23rd, 2012 00:00
thanks for the response. I messed around quite a bit since posting this and eventually I managed to resolve the issue yesterday using the document linked above "Deployment of Dell M-Series blade switches in a Cisco Catalyst Network" <- This is a great read! A lot of my worries about adding power connect into our environment were removed because of this.
I started again with my configs and was much more methodical in working through the process.
Step 1: Create lag between 6224 and 6509, configure trunking and confirm working on something other than VLAN1
Step 2: Create lag between M6348's and 6224, configure as switchport general and tagg traffic
Step 3: Configure Each blade interface as trunks
It really was that simple. I got confused because from the 6224 I could ping everything on our network as long as it wasn't connected directly to the 6224. The reality was I hadn't been explicit enought with the port channel create between the Cisco and powerconnect.