Bug 927349

(In reply to comment #0)
> Steps to Reproduce:
> 1. Create a new instance in openstack
> 2. The leased column in the fixed_ip table will not get updated.
> 3. Delete the new instance in openstack

Please attach options and configuration files that dnsmasq is started with.
As I'm not familiar with setting up openstack I don't want to spend a lot
of time with openstack but rather with fixing dnsmasq. Please also attach
scripts that are supposed to run on startup/when lease is added/when lease
is removed.

Thank you.

/usr/sbin/dnsmasq --strict-order --bind-interfaces --conf-file=/etc/dnsmasq.conf --domain=novalocal --pid-file=/var/lib/nova/networks/nova-br100.pid --listen-address=172.0.0.1 --except-interface=lo --dhcp-range=set:'private',172.0.0.2,static,255.255.248.0,600s --dhcp-lease-max=2048 --dhcp-hostsfile=/var/lib/nova/networks/nova-br100.conf --dhcp-script=/usr/bin/nova-dhcpbridge --leasefile-ro

Created attachment 717182 [details]
nova-dhcpbridge

Created attachment 731497 [details]
Sample testing script

I tested dnsmasq-2.48-13.el6 with my simple testing script (the attached one) on RHEL-6.4. I got the following output:

[root@rhel-64-virtual ~]# dnsmasq -d --strict-order --bind-interfaces --conf-file=/etc/dnsmasq.conf --domain=novalocal --listen-address=192.168.133.2 --except-interface=lo --dhcp-range=set:'private',192.168.133.3,static,255.255.255.0,600s --dhcp-lease-max=251 --dhcp-script=/root/script.sh --leasefile-ro --dhcp-hostsfile=/root/hosts.conf
dnsmasq: started, version 2.48 cachesize 150
dnsmasq: compile time options: IPv6 GNU-getopt DBus no-I18N DHCP TFTP "--bind-interfaces with SO_BINDTODEVICE"
dnsmasq-dhcp: DHCP, static leases only on 192.168.133.3, lease time 10m
dnsmasq: reading /etc/resolv.conf
dnsmasq: using nameserver 192.168.122.1#53
dnsmasq: read /etc/hosts - 2 addresses
dnsmasq: read /root/hosts.conf
dnsmasq-dhcp: DHCPDISCOVER(eth2) 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPOFFER(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPREQUEST(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPACK(eth2) 192.168.133.10 52:54:00:91:cf:62 
SCRIPT CALLED WITH PARAMETERS: add 52:54:00:91:cf:62 192.168.133.10 
dnsmasq-dhcp: DHCPDISCOVER(eth2) 52:54:00:38:7e:6c 
dnsmasq-dhcp: DHCPOFFER(eth2) 192.168.133.20 52:54:00:38:7e:6c 
dnsmasq-dhcp: DHCPREQUEST(eth2) 192.168.133.20 52:54:00:38:7e:6c 
dnsmasq-dhcp: DHCPACK(eth2) 192.168.133.20 52:54:00:38:7e:6c 
SCRIPT CALLED WITH PARAMETERS: add 52:54:00:38:7e:6c 192.168.133.20 
dnsmasq-dhcp: DHCPRELEASE(eth2) 192.168.133.20 52:54:00:38:7e:6c 
SCRIPT CALLED WITH PARAMETERS: del 52:54:00:38:7e:6c 192.168.133.20 
dnsmasq-dhcp: DHCPREQUEST(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPACK(eth2) 192.168.133.10 52:54:00:91:cf:62 
SCRIPT CALLED WITH PARAMETERS: old 52:54:00:91:cf:62 192.168.133.10 
dnsmasq-dhcp: DHCPDISCOVER(eth2) 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPOFFER(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPREQUEST(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPACK(eth2) 192.168.133.10 52:54:00:91:cf:62 
SCRIPT CALLED WITH PARAMETERS: old 52:54:00:91:cf:62 192.168.133.10 
dnsmasq-dhcp: DHCPREQUEST(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPACK(eth2) 192.168.133.10 52:54:00:91:cf:62 
SCRIPT CALLED WITH PARAMETERS: old 52:54:00:91:cf:62 192.168.133.10 
dnsmasq-dhcp: DHCPDISCOVER(eth2) 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPOFFER(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPREQUEST(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPACK(eth2) 192.168.133.10 52:54:00:91:cf:62 
SCRIPT CALLED WITH PARAMETERS: old 52:54:00:91:cf:62 192.168.133.10 
dnsmasq-dhcp: DHCPREQUEST(eth2) 192.168.133.10 52:54:00:91:cf:62 
dnsmasq-dhcp: DHCPACK(eth2) 192.168.133.10 52:54:00:91:cf:62 
SCRIPT CALLED WITH PARAMETERS: old 52:54:00:91:cf:62 192.168.133.10


From my point of view dnsmasq is working properly. Changing Product to Red Hat OpenStack.

For some reasons, it seems to run the script only once when starting dnsmasq.
I've tried replacing the openstack nova-dhcpbridge script by a shell script (like in your test) and it was called only one at startup and then, nothing more happened.

As this pretty much works for everybody else, it would seem that there may be something going on with your environment. What does "nova show" on the instance you created show? Is there any unusual output in any of the log files related to the leased address?

Well, since we're using 2.62 instead of 2.48 it's working fine... also, it appears that you may be right and that the lab we're using has some weird bugs ... but I've just looked at a freshly updated node in our production environment and I see DHCPRELEASE messages !

For some reasons puppet didn't update the package on that node and it's still using 2.48-13 !

I guess you can close this bug report.

Thanks.