DHCP across vlans not working

[Pieter 't Hart]

procurve is not really my "thing", but when looking at :
>>>
vlan 1
name "DEFAULT_VLAN"
untagged 41-48
ip address 172.16.45.1 255.255.255.0
no untagged 3-40,Trk1
<<<
my remark would be :
"no untagged 3-40,Trk1" is not he same as "tagged 3-40,Trk1"
As Trk1 seems like your uplink to another switch, I would expect all vlans "tagged Trk1"

[Pieter 't Hart]

correction,
I overlooked some info.
You say the 3Com is "setup vlan25 only".
Do you mean the 3Com knows about vlan 25?
or do you mean the 3Com is setup default (all ports vlan-1) but connects to the procurve vlan-25?
Then you must set the PVID of Trk1 ports to 25.

[Jeff74]

hostname "ProCurve 2910al-48G Switch"
module 1 type J9147A
ip default-gateway 172.16.45.1
ip routing
vlan 1
name "DEFAULT_VLAN"
untagged 41-48
ip address 172.16.45.1 255.255.255.0
no untagged 1-40
exit
vlan 25
name "lvl2"
untagged 1-2
ip helper-address 172.16.45.10
ip address 172.16.25.1 255.255.255.0
exit
vlan 15
name "lvl1"
untagged 3-40
ip helper-address 172.16.45.10
ip address 172.16.15.1 255.255.255.0
exit
ip route 0.0.0.0 0.0.0.0 172.16.45.254
snmp-server community "public" unrestricted
snmp-server contact "Jeff Powers Ext 237" location "Server Room Lvl 1"
no autorun
password manager
password operator


DHCP server is plugged into port 42 so untagged DEFAULT VLAN 1

DHCP server GW is 172.16.45.1

The uplink is plugged into Port 1 untagged VLAN 25 lvl2 the 3com switch is set to untagged VLAN 25 lvl2 all ports.

Thanks

[AUS Adam]

Hi Jeff,

Looking at the out put from wireshark this is from a successful ipconfig release / renew.

I would get a capture from a failed attempt from the host and the server, the idea is to see where the packet flow fails.

Just a quick one you don't need a 'ip default-gateway' command if you are running 'ip routing'

[Jeff74]

I think I have solved it by turning off spanning tree on the 3com switch which obviously I dont need on a single cable connection between the switches.

I figured this out with watching the packet sniffer and seeing all these spaning tree packets which seemed to be every other packet on the wire.

Will report back tomorrow if this is indeed a solution. Thanks for your help I was going crazy is sometimes hard being a one man band.

Oh I have also turned off the default gateway and all seems happy so things are being fine tuned.

[cenk sasmaztin]

please send me sh tech print

[Sudain]

I wonder if keeping STP on wouldn't be a good idea for you. You never know when an end user will decide plugging a cable into two wall jacks is a good idea...

Though you want to make sure STP on the same version for both switches. I suggest RSTP or MST as STP may be too slow and causing your issues.

[Jeff74]

I do not have spanning tree enabled on either the HP Procurve or 3com switch at present and all is well.

If I renable LACP on 2 ports between the switches for 2GB connectino do I need spanning tree if one of the links drops out?

Thanks

[Jeff Carrell]

Jeff74
May 17, 2010 00:33:28 GMT
If I renable LACP on 2 ports between the switches for 2GB connectino do I need spanning tree if one of the links drops out?
----

No. LACP makes the ports look like one, even for STP.

So, if one link of a LACP group drops and comes back, STP does not know anything occurred.

You see this as well when you create a LACP trunk group, the interfaces become a new interface type of "TRKx".

If you do a 'show span' you will see the interface TRKx at the bottom of the list.

hth...Jeff

[Shadow13]

i think the issue was with the below:

ip default-gateway 172.16.45.1
ip route 0.0.0.0 0.0.0.0 172.16.45.254

each command was pointing to another place, but i think both commands has the same result but pointing to diff. location.

Jeff mentioned that he removed the defualt gateway, that's why it works now.

Not sure if what i mentioned makes sence :D