Failed to launch instances on the compute node

asked 2016-11-07 10:37:43 -0500

praveen gravatar image

updated 2016-11-08 11:21:33 -0500

rbowen gravatar image


nova-scheduler fails to spawn the instances on the compute node and but able to spin up the instances on the main controller node. Always I am getting "No valid host are found. please try again later". there is no error log statements printed on the n-sch.log

2016-11-07 08:00:53.785 ^[[00;36mINFO nova.scheduler.filters.retry_filter [^[[01;36mreq-d95e581e-2027-4517-a4c0-8163d02d317c ^[[00;36madmin demo^[[00;36m] ^[[01;35m^[[00;36mHost [u'cloud02', u'cloud02'] fails.  Previously tried hosts: [[u'cloud02', u'cloud02']]

But cloud02 machine has plenty of resources like 20-cpus, RAM-65GB Harddisk-40GB

edit retag flag offensive close merge delete


nova hypervisor-list ?

dbaxps gravatar imagedbaxps ( 2016-11-07 11:05:45 -0500 )edit

It's showing both hypervisor's running in the controller and compute node

praveen gravatar imagepraveen ( 2016-11-07 12:19:40 -0500 )edit

Upload somewhere :-

1. ifconfig ( both nodes )
2. ovs-vsctl show (both nodes )
3. openstack-status (both nodes)
dbaxps gravatar imagedbaxps ( 2016-11-07 12:32:02 -0500 )edit

The good news is that the scheduler knows about host cloud02, but for some reason it has discarded it. The retry filter is only used at the second and later attempts to find a host, so look for earlier messages about cloud02 in n-sch.log - they should tell you why the scheduler doesn't like it.

Bernd Bausch gravatar imageBernd Bausch ( 2016-11-07 17:15:29 -0500 )edit

Try to check whether the Queue services are working fine in the compute node (rabbitmq). Also check the hypervisor logs under /var/log/libvirt/qemu ( or /var/log/libvirt/libxl case of Xen)

Beny gravatar imageBeny ( 2016-11-08 00:56:11 -0500 )edit