Mike Ruiz wrote: > > Any possibility this is an effect of the goofy network I am in? Or some > corruption from the relay? >
I think that's not a bad insight. Is this effect intermittent? My guess is that the following sequence causes it: 1) Client on remote network gets new lease (DHCP interaction goes via relay) 2) dnsmasq is restarted 3) No new leases are given on the network, so no packets come via the relay 4) The client in step one _renews_ the lease, which will go direct, and not via the relay. In those circumstances, I can see how a cruicial bit of data might not be intialised. A good test would be to find a malfunctioning client and force it to release and then re-aquire a lease. If that works OK we're on the right lines. The main argument I can see against this theory is that it would be amazing that it has not been found up 'till now. Simon. > Here is the dhcpdump of a sample request (not the same as the above): > > --------------------------------------------------------------------------- > TIME: 16:53:48.651085 > IP: > (00:01:e8:76:31:4a) > (00:21:9b:a6:6e:24) > OP: 1 (BOOTPREQUEST) > HTYPE: 1 (Ethernet) > HLEN: 6 > HOPS: 1 > XID: 907bce3f > SECS: 65535 > FLAGS: 0 > CIADDR: 10.32.165.147 > YIADDR: 0.0.0.0 > SIADDR: 0.0.0.0 > GIADDR: 0.0.0.0 > CHADDR: f0:4d:a2:01:62:9f:00:00:00:00:00:00:00:00:00:00 > SNAME: . > FNAME: . > OPTION: 53 ( 1) DHCP message type 3 (DHCPREQUEST) > OPTION: 55 ( 10) Parameter Request List 1 (Subnet mask) > 28 (Broadcast address) > 2 (Time offset) > 3 (Routers) > 15 (Domainname) > 6 (DNS server) > 12 (Host name) > 40 (NIS domain) > 41 (NIS servers) > 42 (NTP servers) > > --------------------------------------------------------------------------- > TIME: 16:53:48.686584 > IP: > (00:21:9b:a6:6e:24) > (00:00:5e:00:01:29) > OP: 2 (BOOTPREPLY) > HTYPE: 1 (Ethernet) > HLEN: 6 > HOPS: 1 > XID: 907bce3f > SECS: 65535 > FLAGS: 0 > CIADDR: 10.32.165.147 > YIADDR: 10.32.165.147 > SIADDR: 172.20.1.116 > GIADDR: 0.0.0.0 > CHADDR: f0:4d:a2:01:62:9f:00:00:00:00:00:00:00:00:00:00 > SNAME: ks.sec.zynga.com. > FNAME: discovery/pxelinux.0. > OPTION: 53 ( 1) DHCP message type 5 (DHCPACK) > OPTION: 54 ( 4) Server identifier 0.0.0.0 > OPTION: 51 ( 4) IP address leasetime 259200 (3d) > OPTION: 58 ( 4) T1 129466 (1d11h57m46s) > OPTION: 59 ( 4) T2 226666 (2d14h57m46s) > OPTION: 1 ( 4) Subnet mask 255.255.255.128 > OPTION: 12 ( 15) Host name netboot-GDNVKN1 > OPTION: 3 ( 4) Routers 10.32.165.129 > OPTION: 42 ( 8) NTP servers 172.20.1.116,172.20.1.124 > OPTION: 6 ( 8) DNS server 172.20.1.117,172.20.1.125 > OPTION: 15 ( 13) Domainname sec.zynga.com > --------------------------------------------------------------------------- > TIME: 16:53:48.712433 > IP: > (00:21:9b:a6:6e:24) > (00:00:5e:00:01:29) > OP: 2 (BOOTPREPLY) > HTYPE: 1 (Ethernet) > HLEN: 6 > HOPS: 1 > XID: 907bce3f > SECS: 65535 > FLAGS: 0 > CIADDR: 10.32.165.147 > YIADDR: 10.32.165.147 > SIADDR: 172.20.1.116 > GIADDR: 0.0.0.0 > CHADDR: f0:4d:a2:01:62:9f:00:00:00:00:00:00:00:00:00:00 > SNAME: ks.sec.zynga.com. > FNAME: discovery/pxelinux.0. > OPTION: 53 ( 1) DHCP message type 5 (DHCPACK) > OPTION: 54 ( 4) Server identifier 0.0.0.0 > OPTION: 51 ( 4) IP address leasetime 259200 (3d) > OPTION: 58 ( 4) T1 118365 (1d8h52m45s) > OPTION: 59 ( 4) T2 215565 (2d11h52m45s) > OPTION: 1 ( 4) Subnet mask 255.255.255.128 > OPTION: 12 ( 15) Host name netboot-GDNVKN1 > OPTION: 3 ( 4) Routers 10.32.165.129 > OPTION: 42 ( 8) NTP servers 172.20.1.116,172.20.1.124 > OPTION: 6 ( 8) DNS server 172.20.1.117,172.20.1.125 > OPTION: 15 ( 13) Domainname sec.zynga.com > --------------------------------------------------------------------------- > > > -Mike > >>> I should mention at this point that the configuration has grown rather >>> large: >>> >>> [root@netops-adm-1 ~]# egrep -v '^#|^$' /etc/dnsmasq.conf | wc >>> 12515 12530 1013088 >>> [root@netops-adm-1 ~]# grep dhcp-host /etc/dnsmasq.conf | wc >>> 5735 5951 514853 >>> >> Nice to see that it keeps going at those sizes: there shouldn't be any >> problems just because of the size of the config. >> >> >> Simon. >> >> >>> -Mike > >