Running OpenStack Icehouse (RDO + EL7) and I noticed an issue today when I rebooted a node hosting cinder volumes. When the node came back online it didn't have any configuration for the iscsi targets.

Doing some research it appears when volumes are created the target configuration is not saved so when the reboot happened the in memory configuration was lost. I'm trying to determine if I've missed a configuration step which would keep this from happening or if there is a bug I'm running into.

[DEFAULT]
rpc_backend = cinder.openstack.common.rpc.impl_kombu
rabbit_host = <snip>.net
rabbit_port = 5672
rabbit_userid = cinder
rabbit_password = <snip>
auth_strategy = keystone
volume_group = cinder-volumes
volume_driver = cinder.volume.drivers.lvm.LVMISCSIDriver
osapi_volume_workers = 48
iscsi_helper = lioadm
iscsi_ip_address = <snip>
glance_host = <snip>.net