Score:0

DHCP load balancing added to production and existing clients

cn flag

I have to ask something I cannot find explained in e.g. https://learn.microsoft.com/en-us/previous-versions/windows/it-pro/windows-server-2012-r2-and-2012/dn338976(v=ws.11) (refers to 2012R2, but that should not matter, my Google-fu failed me)

We have two Windows DHCP servers in our LAN in failover mode, Server A as master and Server B as hot standby. But now we have decided that it makes more sense to run them in load balancing mode. This can easily be acheieved by removing the failover and then creating a new one with different settings.

My question is: What happens to clients who already have an active lease (obtained from Server A while it was master of the failover)? My understanding:

  • Clients will ask for renewal after half the lease time and will do so by unicast with their old DHCP server, i.e., Server A
  • About 50% of these clients will fail because per their hash value, now Server B is responsible for these.

Q1: Will these fail "silently" or will Server A send NACK responses?

Depending on the answer to the preceding question:

  • In case of NACK, the clients will start afresh with DHCPDISCOVER broadcast and will obtain a lease from Server B (probably for a different, "random" address).
  • On the other hand, in case of silent fail, it will take another while until the rebind timer expires, and then the client tries DHCPREQUEST its original IP and it will hopefully be granted by Server B (whom the client will refresh from by unicast in the future)

If my understanding of what happens, it seems to me that silent failure is preferred in mist circumstances, so I hope that it is the answer to my question Q1 and someone can confirm it. Nevertheless, her comes

Q2: Is it possible to configure the silent vs. NACK bahaviour? (And what are possible scenarios where one of the choices would be prefered?)

I sit in a Tesla and translated this thread with Ai:

mangohost

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.