[Dnsmasq-discuss] Strange problem

[Don Muller]

I am having a strange problem that I think might be related to dnssmasq.
Sorry for the length of this but you need the background.

 

I have a nas (QNAP TS-509) that is my main server. It runs linux and a
number of apps one of which is dnsmasq. I use the dns and dhcp features. I
have another nas (QNAP TS-639), which also runs linux,  which I use for
testing and backups. The 639 has a reservation in dnsmasq. This server is
usually off. I have a script on the 509 that wakes the 639, does an rsync,
and then shuts the 639 down. I was testing a new wakenas script when I ran
into a problem. The way the new wakenas script works is like this:

 

1 - ping the 639 to see if it responds.

2 - if it responds

                2a - echo message that the 639 is up.

                2b - exit.

3 - if no response

                3a - issue wake on lan command 

                3b - start loop - loop 21 times

                                4 - ping 639

                                5 - if responds 

                                                5a - echo message that the
639 is up

                                                5b - exit

                                6 - sleep 10

                3c - end loop

                3d - echo message that the 639 is not up

                3e - exit

 

So basically if the 639 is not up it attempts to wake it and waits a max of
3.5 minutes to see if it is up. The script works except that it reports that
the 639 is not up when in fact it did start it. If I ping the 639 by its
name from a terminal session on the 509 it returns an address of
169.254.100.100 when in fact the address of the 639 as assigned by the 509
is 192.168.22.242. If I ping it by its fqdn the correct address is returned.
I think what is happening is that when the 639 is turned on and it first
connects to the network it is connecting with the address of
169.254.100.100. It then issues the dhcp request and gets assigned the
correct address but somehow the old address is still cached somewhere.

 

Some other info about the network. The 509 is also a wins server and the 639
is a wins client. Both the 509 & 639 and two 1gb ethernet connections that
are bound into a single connection using 802.3ad. They are connected to a
Cisco 2960g switch which is configured to bind the ports from each server
into a single 802.3ad trunk.

 

So how can I determine if the problem is with dnsmasq? What logging should I
turn on and what should I be looking for in the logs?

 

Here is the dnsmasq configuration:

 

#             Our log settings

log-facility=/share/MD0_DATA/apps/logs/dnsmasq.log

log-queries

 

#             Our DNS settings

 

domain-needed

expand-hosts

                                

bogus-priv

filterwin2k

resolv-file=/etc/config/user/dnsmasq-resolv.conf

 


#clear-on-reload

stop-dns-rebind

user=xxxxx

group=yyyyyyy

no-hosts              

addn-hosts=/etc/config/user/dnsmasq-hosts.conf

domain=djmuller.com

cache-size=1000               

local-ttl=600


neg-ttl=600

mx-target=mail.djmuller.com

localmx

# address=/adbrite.com/127.0.0.1


 

#             Our dhcp settings

 

dhcp-range=192.168.22.1,192.168.22.50,24h

dhcp-option=vendor:MSFT,2,1i

dhcp-leasefile=/share/MD0_DATA/apps/dnsmasq/dnsmasq.leases

dhcp-authoritative

dhcp-hostsfile=/share/MD0_DATA/apps/config/dnsmasq/dnsmasq-dhcp-reservations
.conf

dhcp-optsfile=/share/MD0_DATA/apps/config/dnsmasq/dnsmasq-dhcp-options.conf

 

#             ad servers contained in additional configuration file

 

conf-file=/share/MD0_DATA/apps/config/dnsmasq/dnsmasq-adservers.conf   #
just contains address=/domain/ip entries for ad servers

 

 

I hope I have explained my problem so that you understand it.

 

Thanks

Don

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.thekelleys.org.uk/pipermail/dnsmasq-discuss/attachments/20100510/4b6df6d2/attachment.htm