DHCP/CARP

scorpionsforgotgmail.com

Good Morning,
Continuing the investigation of my problem in dhcp / carp. I managed to solve part of it. However, I noticed that the behavior is not correct. When the dhcp master is enabled and the slave too (after the slave restart), I see that for a few minutes (see image below) the funcinamento is normal, the message on the slave recover-wait is displayed correctly and the partner-down master . The problem is that after a few minutes this setting is changed. Both DHCPs are like My stare normal (See picture). I do not know what could be, since for a few minutes the two usually work in sync. Anyone know what to do to solve this?

Normal

Thanks

normal1.png_thumb
![not normal.png](/public/imported_attachments/1/not normal.png)
![not normal.png_thumb](/public/imported_attachments/1/not normal.png_thumb)

Derelict

normal/normal is what you want.

You have received the same answer on the mailing list already:

scorpionsforgot@gmail.comwrote:
Good Morning

The dhcp in secondary carp is even distributing IP with the active
master . Anyone know how to solve this ?/scorpionsforgot@gmail.com

It's not a problem, that's how it's supposed to work.

That's from cmb. You can consider it correct.

scorpionsforgotgmail.com

Sory, I did not understand.

This can not be normal behavior. There can be two DHCPs network distributing IP, the slave may even become active, but can not deliver IP. If you look at the images, the first image shows it as it should be. In the second image, the two servers are as normal state. And both deliver ip. This can not, should not happen.

Derelict

No. The first image shows it as it is in recovery mode after a failure.

The second, the one that says normal/normal is how it is supposed to be. That's why the state of both nodes is normal. Look up the English word normal.

It is by design. They split the pool and keep track of the leases each node has assigned.

scorpionsforgotgmail.com

derelict
Thanks for your answer. From what you said, right then he would be with normal status in both and not as in the first image. OK! Now the doubt more important if this state "Normal " for the two DHCPs when this state are distributing IP , and another image with status recover -wait just a delivery . We can not have two DHCPs distributing IP on the same network , sorry , for me this behavior may not be correct . If in the normal state , and only the master was distributing IP , ok , but the two distributing ip , this is wrong . What do you think is the problem?

Derelict

You can when they are working together by design.

piwwo

I think that I actually have the same issue: Both servers are providing an IP and a route to a request but only the master should (or the failover backup when the master is not responding)

This question seems to be from a Windows-Environment and not ISC dhcp but the behaviour should be the same, or?

DHCP Failover partnerships create a hash of the client MAC address when a lease request is recieved. Both servers will know which hashes it needs to respond to and both servers should be able to see requests. If a request goes unanswered after a handful of tries from the client, the partner server will respond with a short term lease from its IP pool equal to the MCLT length. (So it can expire the lease quickly when the unavailable server is back up).

http://serverfault.com/questions/614331/dhcp-failover-didnt-work-when-one-server-was-offline

Derelict

Pretty much all I can say is if you do not like the way ISC DHCPD works (and therefore pfSense as a DHCP server), use a different DHCP server.

What is the problem you are seeing that you are trying to solve?

Both nodes work together by design to serve the DHCP load. They cooperate as a unit. It is not the same thing as having a second, or rogue, DHCP server on your network.

piwwo

No, I am not complaining about ISC, I know the server is widely used in many Linux/Unix environments. I just wanted to know if this was a missconfiguration on my side or normal behaviour.