DHCP problem: instances cannot obtain a specific IP

asked 2014-02-27 04:28:33 -0600

pernacentus gravatar image

updated 2014-02-27 14:59:07 -0600

smaffulli gravatar image

I'm currently running Openstack Havana with Neutron on my infrastructure. Since yesterday, any booted instance could correctly obtain an IP address from the DHCP server. Yesterday afternoon I've terminated an instance with IP and now, if I try to boot another instance with the same IP, it fails obtaining the address.

However, if I try to boot any other instance with a different IP, it correctly gets an IP address. I've found that dnsmasq mantains a list of mac/IP pairs in the file /opt/stack/data/neutron/dhcp/d333453f-d5d5-4077-b33a-85ec42359422/host. According to this post this file is modified by neutron-dhcp-agent everytime a port has been created. In my own file there is the following line:


but this mac address does not correspond to any port in my infrastructure (I've also checked any table in neutron_ml2 database). Every time I try to boot a new instance with IP, a new line is appended to the file with the same IP but a different mac address, but the former line prevents the newly booted VM from getting an IP. I've also tried to manually delete the line, but Neutron still continues to write it in the file. Is there anybody that have experienced this issue?

edit retag flag offensive close merge delete