01-11-2015 02:51 AM - edited 03-07-2019 10:10 PM
Hello there,
I have this dilemma in my network. Each time we have power outage (It happens a lot and we don't have a UPS yet) some client IPs go crazy.
The core is the DHCP Server for several pools and VLANs. I have 10s of PCs attached to the PTS switch and I will concentrate on just two of them for now. Here is an expert from the runing-config file on the Core Switch:
!
ip dhcp pool PTS
network 10.100.10.0 255.255.255.0
default-router 10.100.10.254
dns-server 192.168.1.100 192.168.1.101
domain-name example.com
lease 21
!
and PCs are told to automatically assign an IP address from a DHCP Server.
Scenarios 1 - Normal:
PC1 IP: 10.100.10.11
PC2 IP: 10.100.10.22
Scenarios 2 - After Power Outage on PCs Only:
PC1 IP: 10.100.10.11
PC2 IP: 10.100.10.22
Scenarios 3 - After Power Outage for ALL:
PC1 IP: 10.100.10.11
PC2 IP: 10.100.10.33
As far as I know, the IP should not change before the lease time expires, but why does that happen only to PC2 in scenario 3? What could be the problem? it looks like that PC2 is requesting a new IP instead of renewing the already existing one. Can unsynchronized timing be an issue? It is really annoying as we started to have blackouts very often nowadays and the UPS is not here yet .
Help Is highly appreciated.
Thank you.
Ahmed.
Solved! Go to Solution.
01-11-2015 08:35 AM
First of all, configure a database agent for DHCP:
http://www.cisco.com/en/US/docs/ios/12_4t/ip_addr/configuration/guide/htdhcpsv.html#wp1061906
01-11-2015 07:31 AM
Hallo,
by power outage for all you mean the DHCP-Server (switch) and the clients are without power? If that is the case then who is supposed to remember the old lease? I have seen printers that did remember their old IP even when shut off, however this behaviour is nothing you can really rely on.
If you want to make sure all PCs always get the same ip address you need to configure static dhcp bindings. However when you start doing this you will most probably realize that a cisco switch is not the best option as DHCP server.
IIRC it is possible to configure the switch in a way that it is saving the current DHCP leases to a TFTP-File and reload them after reboot, but that is certainly a fiddly solution.
Best practice would be to get a dedicated DHCP server, ISC or windows or whatever front end you prefer.
01-11-2015 10:33 PM
Yes I mean the DHCP-Server (switch) and the clients are without power. Now the question is: where is the database stored? if in RAM, then how did PC1 store its original IP address in scenario 3? if in NVRAM, then WHY did PC2 lose its original IP address in the same scenario?
Thank you :)
01-11-2015 08:35 AM
First of all, configure a database agent for DHCP:
http://www.cisco.com/en/US/docs/ios/12_4t/ip_addr/configuration/guide/htdhcpsv.html#wp1061906
01-11-2015 10:32 PM
I will try and do that!
Thank you :)
01-11-2015 11:57 PM
Some good advice from others there.
If the issue carries on then it may be worth doing a packet capture on one of the devices when the issue occurs. Look at the DHCP packets in wireshark and see if they are doing a Discover or simply a Request and asking for a specific IP address.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide