Wired Intelligent Edge

I have an HP ProCurve 2520g that recently started having an issue where it will only gives access to the network if the PC's have a static IP, Gateway and DNS server entries. THe PC's get a DHCP address, but cannot get access to the network.  I updated the firmware to the latest version...but a few of the PC's connected to it still have the same issue...some are better.  This has to do something with this switch as all other switches are functioning fine, and if I take the PC's and plug them into another switch they are then fine as well...they grab a DHCP address and access the network fine.  Any ideas of where to start.  No config changes have been made.  It was literally all of a suddent the PC's lost connection to the network and the only way to get them back on was by static IP.

 

Thanks

#DHCP

Hi Moondog,

 

Sounds like to me the switch has had a bit of a paddy.. I would suggest a factory reset and then upload the running-config back to it, if this doesn't fix it I would say it's a call to HP support.

Hi moondog3440,

If the switch works fine when the PCs have a static IP, i very much doubt that it is faulty, but it's difficult to be sure with the amount of detail you've provided.  Where is your DHCP server?  Was the switch set up as a DHCP relay?  What VLANs do you have?  How are they routed?  If you provide us with more info we can probably help track it down further.

The switch works when I set a PC to static IP, but the next day they come in I have to go and change the IP to another static IP for it to work correctly.  Its very odd.  I have 3 other PC's that are completely fine connected to the same switch.  Here is the running config of the switch:

 

hostname "BO-2520G-PoE"
time timezone -6
qos dscp-map 000000 priority 0
qos dscp-map 001000 priority 1
qos dscp-map 010000 priority 2
qos dscp-map 011000 priority 3
qos dscp-map 100000 priority 4
qos dscp-map 101000 priority 5
qos dscp-map 110000 priority 6
qos dscp-map 111000 priority 7
interface 5
no power-over-ethernet
exit
ip default-gateway 10.9.1.1
vlan 1
name "DEFAULT_VLAN"
untagged 24
no untagged 1-23
no ip address
exit
vlan 20
name "Voice"
tagged 1-24
voice
no ip address
exit
vlan 100
name "Management"
ip address 10.9.1.30 255.255.0.0
tagged 24
exit
vlan 40
name "Wireless"
tagged 24
no ip address
exit
vlan 11
name "BO Users"
untagged 1-23
tagged 24
no ip address
exit
snmp-server community "fpbsdpub" operator unrestricted
primary-vlan 100
management-vlan 100

 

Hi moondog3440,

My spidey sense is telling me that this is nothing to do with the switch at all. What are the IP addresses of your working and non-working devices? Which ports are they plugged into?

Why would this matter though....no config changes were made and no changes as to where the PC's plug into changed...the switch just all of a sudden just started having this issue??? This switch has been functioning normally for a little over a year.

It might still be a faulty switch, as Chris suggested.  It just doesn't sound that way to me; it sounds like something else in your network has changed.

 

If you get a clear problem description, with data about exactly which ports and IPs work and which ones don't, we'll be well on the way to solving the problem.

port 9 doesn't work using DHCP and I have to change static IP's almost daily and 7 does work on DHCP as normal.  As far as which IP's work..... it depends on the PC....on the PC's that are ok any DHCP address will work.  On a PC that is not ok no IP will work as DHCP or an IP that has been used before.

 

But I can take those PC's that don't work....plug them into a different switch and they work off of DHCP just fine.

I can't help but think we're still missing some things here.  Where is your DHCP server?  What OS does it run?  What is the dynamic pool?  What addresses are currently listed as leases?  Are there reservations for selected hosts?  If you put in a reservation for your test system, does it work reliably?  Are there any DHCP log messages which indicate issues?

 

As well as Pete's suggestion about using wireshark, i would consider installing arpwatch (or a similar utility if you use Windows servers) to check for duplicate IPs and MAC<->IP  pair flapping.

The switch is on the 10.9.x.x subnet and the DHCP server is on our 10.1.x.x subnet.  The DHCP server is on a Windows 2008 r2 sp1 box.  The address pool is 10.6.40.1 - 10.6.41.254.  Current leases 10.6.40.5 and 10.6.40.8.  No reservations.  There are no duplicate addresses either.  I have not tried to make a reservation at all.  Ther are no DHCP log messages either that would indicate an issue.  I had alo updated the server that runs DHCP and that did not resolve the issue either.  To add a wrinkle to this now I have static printers that cannot be pinged by PC outside of this switch but they can buy PC's that are serviced by this switch.  However I can ping the two PC's mentioned above 10.6.40.5&8 from outside the switch.

 

WEIRD.......

Hi moondog3440,

If i'm understanding your config correctly, 10.6.40.0/23 (VLAN 11) is where your clients are not working (ports 1-23 untagged).

If that's the case, this switch has nothing to do with DHCP, since it has no address on VLAN 11, routing is not turned on, and there is no ip helper defined. You need to investigate this issue in the switch(es) higher up the chain, starting with the one that this one connects to on port 24.

I think you need to break out Wireshark and see whether the DHCP packets are being sent/recieved at each end.

Pete W