Odhcpd DHCPv6 server problem after a service restart

[Livy]

Hi everyone.

For some reason my Windows 11 hosts failed to extend lease time or get addresses after odhcpd service restart. Not all are affected, but I can say there are a lot.

I cannot pinpoint the exact cause right now, but it is very noticable if you have 300-400 clients and a short lease time like 1h. Symtom: after the lease time expires, the clients are left with no ULA leases (I have not tested with GUA yet, but it is probably the same). Without SLAAC, hosts lose IPv6 routing and people immediately complain. RA continues to work, as I can see the default gateway and DNS server being set correctly, albeit with the useless LLAs.

To make them work without unplug/replug the network cable, I have to issue the command ipconfig /renew6. And then they work, which means odhcpd DHCPv6 function is working. But I cannot explain why the old leases, before the service odhcpd restart, just expire and my Windoze clients stand there listening to RAs without even trying to start over with a DHCP SOLICIT.

I know using SLAAC in conjunction with DHCPv6 is a safety measure, but it is out of this question.

Any thoughts?

P.S: well, I just realize that the leasefile '/tmp/odhcpd.leases' gets wiped when the odhcpd service is restarted. What is going on here?

[JustAnotherEndUser]

@Livy

I just realize that the leasefile '/tmp/odhcpd.leases' gets wiped when the odhcpd service is restarted. What is going on here?

From what I could find, the lease file is write-only. The previous lease file does not get read at service startup, but overwritten instead. It’s purpose is to relay lease info, such as to display in LUCI ‘Overview’ page, or to be read by a script, etc.

This is an older thread, but has related information:

Odhcpd | How to make lease file persistent?

Hi! I moved the odhcpd lease file to a permanent storage. But the lease file gets reset/deleted every reboot. How to make the lease file persistent/survive a reboot? Thanks.

Not sure about this, but possible workaround?

[solved] Stale DHCPv6 leases under Network -> DHCP and DNS

one way: use /etc/ethers ( and related dnsmasq option ) #readethers boolean 0 -Z Read static lease entries from /etc/ethers, re-read on SIGHUP another way: adapt something like this to populate dnsmasq A records; ( excerpt from ipv4 script ) #!/bin/sh while read timeS maC iP namE maC2; do echo "$iP $namE" #ADDandorREMOVE A RECORD or SETUP STATIX-RESERVO sleep 1 done < dhcp6.leases #(/var/hosts/odhcpd) service reload exit #schedule every 5-10mins in cron or trigger in /etc/dnsmasq.conf…

[optilumin0x1]

Isnt the RA server still needed for telling user device where to get ip ? I remember there is specific flags (m,o,h) on RA option page and you need to set it to M in order for client device to registered themselves on dhcpv6

[Livy]

Well. Some more debug info:

root@OpenWrt:~# cat /tmp/odhcpd.leases | wc -l
200
root@OpenWrt:~# service odhcpd restart
root@OpenWrt:~# cat /tmp/odhcpd.leases | wc -l
0
root@OpenWrt:~# cat /tmp/odhcpd.leases | wc -l
10
root@OpenWrt:~# cat /tmp/odhcpd.leases | wc -l
36

So a restart shall wipe the lease file. Is it intended?