Bug #898290: iSCSI volume backend treats FLAGS.host as a hostname
Change ISCSIDriver to set volume.provider_location during create_export. This records the location of the LUN, so that nova-compute does not need to run the iSCSI discovery code itself. As part of this, include the IP address of the target (--iscsi_ip_address) in the provider_location. This means that we don't use volume's host identifier (which could be an opaque ID) when trying to connect to the iSCSI target -- we use the admin-specified IP address instead. The string-join to set provider_location is shared with ZadaraBEDriver, which was doing something similar. I've brought that into a helper function. The docstring for the --host flag has been clarified. Change-Id: I8402da86345e786a46a4d222ad4d8a4449d2bd3f
This commit is contained in:
@@ -440,7 +440,8 @@ DEFINE_string('image_service', 'nova.image.glance.GlanceImageService',
|
||||
'The service to use for retrieving and searching for images.')
|
||||
|
||||
DEFINE_string('host', socket.gethostname(),
|
||||
'name of this node')
|
||||
'Name of this node. This can be an opaque identifier. It is '
|
||||
'not necessarily a hostname, FQDN, or IP address.')
|
||||
|
||||
DEFINE_string('node_availability_zone', 'nova',
|
||||
'availability zone of this node')
|
||||
|
Reference in New Issue
Block a user