Glance and Cinder with same NFS Backend

asked 2017-01-21 06:11:42 -0500

theSuess gravatar image

I recently installed a fresh Openstack Cluster as described in the official Documentation (CentOS based)

I have a working cluster with one controller, 4 compute nodes and cinder running on the controller node. Cinder is backed by a NAS via a NFS share and the glance images are stored locally on the controller node. Everything works fine as long as I use small (<1GB) images from Glance. Everything larger results in a nova error "Block device mapping is invalid". The error in the logs look like a timeout error as the state is set to "downloading"

2017-01-21 11:54:13.929 3245 ERROR nova.compute.manager [instance: 8024cf2d-a45a-4a3a-a714-fb4b3f6e20c5] VolumeNotCreated: Volume c9ba08f6-442c-4a8c-a0b4-f95b18876cb7 did not finish being created even after we waited 191 seconds or 61 attempts. And its status is downloading.

My question is: how can I optimize glance and cinder to keep my data on the NAS and stop nova from pulling the entire image. Even a pointer in the right direction would help! Thanks!

edit retag flag offensive close merge delete