[Openstack] VM cannot get IP in Big cluster (grizzly, quantum, multi-host, gre)

Yuan, Ye A ye.a.yuan at intel.com
Thu Jul 25 12:04:53 UTC 2013


I also met this bug:
http://openstack.redhat.com/forum/discussion/230/warning-quantum-db-agentschedulers_db-fail-scheduling-network/p1

Best wishes,
Yale Yuan

From: Openstack [mailto:openstack-bounces+ye.a.yuan=intel.com at lists.launchpad.net] On Behalf Of Yuan, Ye A
Sent: Thursday, July 25, 2013 7:58 PM
To: openstack at lists.launchpad.net
Subject: Re: [Openstack] VM cannot get IP in Big cluster (grizzly, quantum, multi-host, gre)

Hi all,
Our team has built 15+ compute nodes Openstack cluster, and with 200+ VM running on it. But we met a stranger issue and boring me several work days. Could anyone help me ? :)
The phenomenon is: if we create multi VMs in one request some or all of them cannot get IP from DHCP agent. Logs in quantum-server report "no active DHCP agent" .I read official doc and enabled DHCP HA, but all DHCP agents for network are inactive while instance boot and suddenly change to active. "no active DHCP agent" still occurs. In other words, the agent heartbeat always go active and immediately go inactive. I set agent_down_time = 120 but does not work :(
My quantum server log is very big and if anyone interesting this issue , I will send it to you.
Openstack is Grizzly, quantum using GRE mode, multi-host is enabled.

Best wishes,
Yale Yuan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack/attachments/20130725/b7629a01/attachment.html>


More information about the Openstack mailing list