provider-instance Error from libvirt during destroy

asked 2018-10-22 14:06:35 -0500

New openstack rocky install. Followed the self-service network model. All servers are virtual machines under kvm. Installed controller node (Cinder+Horizon+Neutron+Nova+Glance+Keystone) Compute1 node(Nova+Neutron) Block node(cinder) Defined provider-instance on the provider network. Can ping and ssh to it. Defined selfservice-instance on the selfservice network and associated floating ip. Can ping and ssh from outside. Shutting down selfservice-instance proceeds with no issues. When shutting down provider-instance it stays up. Nova logs show Error from libvirt during destroy. Code=38 Error=Failed to terminate process 2109 with SIGKILL: Device or resource busy; attempt 1 of 3 Its retried 3 times and eventually fails with Exception during message handling: libvirtError: Failed to terminate process 2109 with SIGKILL: Device or resource busy Lateron when shutting down the compute node the console libvirt also reports it's trying to shutdown a guest for a number of retries until eventually times out and shutdown proceeds. Would appreciate any ideas on how to narrow the issue. Have tried with the provider-instance having cinder volumes attached and without them. Mades no difference. Thanks a lot

edit retag flag offensive close merge delete