Re: dhcpd synchronization: leases recovery after downtime
Hello Edgar, Le 19/07/2020 à 00:35, ed...@pettijohn-web.com a écrit : > Not sure how to integrate it properly but this sounds like a job for > ifstated(8). Triggering a sync at interface up status change wouldn't be enough: for instance if only dhcpd crashes and recovers later, there is no change on interface status.
Re: dhcpd synchronization: leases recovery after downtime
Hello Janne, Le 19/07/2020 à 10:21, Janne Johansson a écrit : > 2) dhcpd checks that ip's don't reply to ping (or exist in arp?) before > handing out an ip from a dynamic range How did you check that? I'm not as sure as you about that. For instance if you define a static lease for a given machine, you have to exclude the IP address from your dynamic range. So I don't think there is any kind of check if the address is free on the network before assigning it. There is no such piece of information in the man page either. Also, even if it worked, I would like to know the purpose of the lease sync mechanism; indeed, because of the lack of downtime recovery, high availability isn't the reason of its existence.
Re: dhcpd synchronization: leases recovery after downtime
Den lör 18 juli 2020 kl 23:28 skrev Guy Godfroy : > Hello, > > I am using two routers on OpenBSD (called mulder and scully), and I wish > to make dhcpd listen on a carp interface between both of them. I am > using the synchronization mechanism: > I noticed the same issue long time ago, but settled for just running two unconnected dhcpds and made sure that 1) all fixed replies exist on both (and clients don't mind getting two answers, they pick the first and stop listening for any extra replies) and 2) dhcpd checks that ip's don't reply to ping (or exist in arp?) before handing out an ip from a dynamic range and this seems to cover most of my concerns, no client would get a different offer from both dhcpds and ack both, and putting as many fixed entries as possible on important hosts to make sure they would work in any case. -- May the most significant bit of your life be positive.
Re: dhcpd synchronization: leases recovery after downtime
On Jul 18, 2020 3:25 PM, Guy Godfroy wrote: Hello, I am using two routers on OpenBSD (called mulder and scully), and I wish to make dhcpd listen on a carp interface between both of them. I am using the synchronization mechanism: mulder# rcctl get dhcpd flags -y em2 -Y em2 carp0 scully# rcctl get dhcpd flags -y em2 -Y em2 carp0 I tested it and it works great when both routers are up. But what is explained in the man page is that it is only a realtime sync. So if I understand correctly, when one of the hosts is down, there is no way for it to get the new leases when it gets back up. My thought about that is that it defeats the original purpose of the feature. My concerns are confirmed by the test: 1. Make sure that carp0 master is mulder. 2. Stop dhcpd on scully. 3. Plug a new machine using a dhcp client on the network of carp0. 4. Observe that the lease is present in /etc/dhcpd.leases on mulder. 5. Start dhcpd on scully. 6. Observe that the lease is missing in /etc/dhcpd.leases on scully. I have found no information about leases synchronization other than the one given in dhcpd man page, but I guess there is no possible leases recovery from other routers after a downtime, or when a new router is inserted in the stack. What is the correct solution to manage this case? Is there a possibility for a router to ask for its missing leases when it gets up? Something like this: 0. mulder is up, scully is down. 1. scully gets up and asks on multicast: "My last lease was created on date d, what's up since that date?" 2. mulder answers: "Here is the leases created since d: ..." Thanks. Guy Not sure how to integrate it properly but this sounds like a job for ifstated(8). Edgar
dhcpd synchronization: leases recovery after downtime
Hello, I am using two routers on OpenBSD (called mulder and scully), and I wish to make dhcpd listen on a carp interface between both of them. I am using the synchronization mechanism: mulder# rcctl get dhcpd flags -y em2 -Y em2 carp0 scully# rcctl get dhcpd flags -y em2 -Y em2 carp0 I tested it and it works great when both routers are up. But what is explained in the man page is that it is only a realtime sync. So if I understand correctly, when one of the hosts is down, there is no way for it to get the new leases when it gets back up. My thought about that is that it defeats the original purpose of the feature. My concerns are confirmed by the test: 1. Make sure that carp0 master is mulder. 2. Stop dhcpd on scully. 3. Plug a new machine using a dhcp client on the network of carp0. 4. Observe that the lease is present in /etc/dhcpd.leases on mulder. 5. Start dhcpd on scully. 6. Observe that the lease is missing in /etc/dhcpd.leases on scully. I have found no information about leases synchronization other than the one given in dhcpd man page, but I guess there is no possible leases recovery from other routers after a downtime, or when a new router is inserted in the stack. What is the correct solution to manage this case? Is there a possibility for a router to ask for its missing leases when it gets up? Something like this: 0. mulder is up, scully is down. 1. scully gets up and asks on multicast: "My last lease was created on date d, what's up since that date?" 2. mulder answers: "Here is the leases created since d: ..." Thanks. Guy
