Description of problem: virt-v2v only checks for disks attached to devices with ResourceType 5 or 6 (an IDE Controller and a Parallel SCSI HBA). Disks attached to any other type of controller (FC HBA, iSCSI HBA, IB HCA, Other storage device) will not be picked up. Version-Release number of selected component (if applicable): Fedora 20/virt-v2v 0.9.0 How reproducible: Attempt to import an OVA with an OVF with any disks attached to a non-IDE or parallel SCSI HBA. Disks attached to those controllers will not be detected. Additional info: The OVF standards specify CIM_ResourceAllocationSettingData (http://schemas.dmtf.org/wbem/cim-html/2.39.0+/CIM_ResourceAllocationSettingData.html) as ruling standard for ResourceType. According to this spec, 5 and 6 are IDE Controller and Parallel SCSI HBA, resp. For example VirtualBox's SATA controller will therefore (for lack of an ID for SATA controllers) become ID 20, Other storage device. The same would logically follow for SAS controllers, while FC, iSCSI and IB controllers have their own ID's. virt-v2v doesn't consider any of these ID's and therefore can't pick up on any of those disks. Another interesting point is that it also doesn't keep the controller metadata, instead collapsing into 'hd[a-z]' and 'sd[a-z]' disks.
Created attachment 853903 [details] This proposed patch makes the device list OVF-compliant, checks all possible controllers for disks and preserves bustypes Proposed patch: * hw_families & naming are now OVF compliant * All listed controllers (the list is easily modifiable) are checked for disks * Disks (attempt to) have their bustype set correctly by an additional metadata type and by evaluating the controller's ResourceSubType * Both DVD's and CD-ROM's are now mapped in as cdroms
Hi, thanks for the bug report. Unfortunately it is filed against a very old version of virt-v2v (0.9) which we don't support at all. Luckily there is a newer version which almost certainly fixes your bug already. It's available starting in Fedora 21. If you still experience this bug, please try with the new version, and reopen if it still happens.
(In reply to Richard W.M. Jones from comment #1) > Unfortunately it is filed against a > very old version of virt-v2v (0.9) Well, that tends to happen if you wait 16 months with a reply to a ready2commit submitted patch.