[Dnsmasq-discuss] randomly stops resolving and handling dhcp requests

[simon-dnsmasq at uc.org]

Hi,

I'm having an intermittent issue where dnsmasq stops responding to DNS and DHCP
requests and does not recover until I kill the process and restart. It can go
for weeks without surfacing, and then happen repeatedly in the span of a few
hours. I have a bridge, br0, which is between a wired and wireless network, no
firewall rules on it.

Any ideas how I might track down the trigger, or more useful information I
should be grabbing when it occurs?

Cheers,
Simon P. Ditner

Config:
------

kernel 2.6.20-16-generic
dnsmasq version 2.41

/etc/dnsmasq.conf:
domain-needed
bogus-priv
interface=br0
dhcp-range=192.168.100.100,192.168.100.254,24h
dhcp-lease-max=50
dhcp-authoritative
dhcp-mac=linksys,00:0E:08:*:*:*
dhcp-option=linksys,66,"192.168.100.1"

Failures:
--------

Example 1:

strace:
15:53:07.351736 time(NULL)              = 1213213987
15:53:07.351927 stat64("/etc/resolv.conf", {st_mode=S_IFREG|0644, st_size=58, ...}) = 0
15:53:07.352246 recvmsg(4, {msg_name(16)={sa_family=AF_INET, sin_port=htons(60831), sin_addr=inet_addr("192.168.100.60")}, msg_iov(1)=[{"\266\236\1\0\0\1\0\0\0\0\0\0\nbannerfarm\3ace\vadve"..., 1280}], msg_controllen=24, {cmsg_len=24, cmsg_level=SOL_IP, cmsg_type=, ...}, msg_flags=0}, 0) = 48
15:53:07.352640 ioctl(4, SIOCGIFNAME, {ifr_index=17, ifr_name="br0"}) = 0
15:53:07.352909 sendto(11, "\376\203\1\0\0\1\0\0\0\0\0\0\nbannerfarm\3ace\vadve"..., 48, 0, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("64.71.255.198")}, 16) = 48
15:53:07.353317 select(12, [3 4 5 6 8 11], [], [], NULL) = 1 (in [11])
15:53:07.362387 time(NULL)              = 1213213987
15:53:07.362578 recvfrom(11, "\376\203\201\2
15:55:00.845955 +++ killed by SIGKILL +++

dnsmasq: DHCPDISCOVER(br0) 192.168.100.203 52:54:00:12:34:56 
dnsmasq: DHCPOFFER(br0) 192.168.100.201 52:54:00:12:34:56 
dnsmasq: DHCPREQUEST(br0) 192.168.100.201 52:54:00:12:34:56 
dnsmasq: DHCPACK(br0) 192.168.100.201 52:54:00:12:34:56 gateway
dnsmasq: not giving name gateway to the DHCP lease of 192.168.100.201 because the name exists in /etc/hosts with address 127.0.1.1

Example 2:

strace:
16:33:08.572298 time(NULL)              = 1213216388
16:33:08.572501 recvfrom(11, "\n\320\201\200\0\1\0\4\0\t\0\t\7content\10worldnow\3co"..., 1280, 0, {sa_family=AF_INET, sin_port=htons(53), sin_addr=inet_addr("64.71.255.198")}, [16]) = 453
16:33:08.572924 time(NULL)              = 1213216388
16:33:08.573083 stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1252, ...}) = 0
16:33:08.573391 stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1252, ...}) = 0
16:33:08.573673 stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1252, ...}) = 0
16:33:08.573971 writev(2, [{"dnsmasq: reply content.worldnow."..., 46}, {"\n", 1}], 2) = 47
16:33:08.574234 send(10, "<135>Jun 11 16:33:08 dnsmasq: re"..., 67, MSG_NOSIGNAL) = 67
16:36:11.278393 +++ killed by SIGKILL +++

dnsmasq: reply a1528.g.akamai.net is 24.153.19.145
dnsmasq: reply a1528.g.akamai.net is 24.153.19.154
dnsmasq: query[A] content.worldnow.com from 192.168.100.60
dnsmasq: forwarded content.worldnow.com to 64.71.255.198
dnsmasq: reply content.worldnow.com is <CNAME>

Example 3:

strace:
17:21:15.252922 time(NULL)              = 1213219275
17:21:15.253079 stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1252, ...}) = 0
17:21:15.253382 stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1252, ...}) = 0
17:21:15.253659 stat64("/etc/localtime", {st_mode=S_IFREG|0644, st_size=1252, ...}) = 0
17:21:15.253948 writev(2, [{"dnsmasq: reply thekelleys.org.uk"..., 49}, {"\n", 1}], 2) = 50
17:21:15.254278 send(10, "<135>Jun 11 17:21:15 dnsmasq: re"..., 70, MSG_NOSIGNAL) = 70
17:21:15.254704 sendmsg(4, {msg_name(16)={sa_family=AF_INET, sin_port=htons(34573), sin_addr=inet_addr("192.168.100.60")}, msg_iov(1)=[{"\36\236\205\200\0\1\0\2\0\5\0\0\3www\nthekelleys\3org\2"..., 171}], msg_controllen=24, {cmsg_len=24, cmsg_level=SOL_IP, cmsg_type=, ...}, msg_flags=0}, 0) = 171
17:21:15.255239 select(12, [3 4 5 6 8 11], [], [], NULL) = ? ERESTARTNOHAND (To be restarted)
17:21:15.904756 +++ killed by SIGKILL +++

dnsmasq: reply www.l.google.com is 72.14.205.104
dnsmasq: query[A] www.thekelleys.org.uk from 192.168.100.60
dnsmasq: forwarded www.thekelleys.org.uk to 64.71.255.198
dnsmasq: reply www.thekelleys.org.uk is <CNAME>
dnsmasq: reply thekelleys.org.uk is 86.14.238.176