Created attachment 1425772 [details] The error from ovirt-engine.log Description of problem: I recently broke my oVirt installation. The mailing list helped me do a fresh re-install. Once that was up and functioning correctly, I imported my VM's from the previous install. All but one imported correctly and appear to be functional without issue. The one throws an error and the list asked that I open a report. When I try to import that one VM I get the error "General command validation failure." within the GUI. I have added as an attachment the output of the ovirt-engine.log. Version-Release number of selected component (if applicable): 4.2.2.2 How reproducible: For me, every time. Steps to Reproduce: 1. Storage -> Domains -> Data Domain -> VM Import 2. Select the VM -> Import 3. Click OK in the pop-up window. Actual results: "Error while executing action: Infra01: General command validation failure." Expected results: An imported VM like all the rest. Additional info: OS: Scientific Linux 7.4 fully updated. For self-hosted engine and the 4 hypervisor hosts. Storage: NFS mount. Roy Golan posted to the list the following information (I am not sure how to interpret it so I hope it is useful - it is just the best lead I've got right now): "I suspect List<Guid> guids = Guid.createGuidListFromString(snap.getMemoryVolume()); where guids is empty because there is not memory volume. which fails this StorageDomain sd = getStorageDomainDao().getForStoragePool(guids.get(0), params.getStoragePoolId()); " What else can I give you to help with the issue?
I selected the correct version in the menu, typed the wrong one in the comment. My apologies. The version is 4.2.2.6.
Tal, looks like you/Maor touched it recently
Can you reproduce this bug?
Maor said on thread: """ It seems like a regression from 55a71302ef91c91e8d9d418794ea3be990dcfe71 Since memory disk was not considered a disk entity in the engine, the VM regisatration worked well. Now, that the memory disk became an entity and the validation was added, it should also be addressed in the VM's OVF. Now it isn't part of the VM's OVF and the disk does not get created in the DB and that is probably why the operation fails. """
Natalie, pease check, thanks
I managed to reproduce it myself on 4.2.z, working on a solution
Pushed a fix that handles the NPE yet there's a major flow in the flow that needs to be fixed, I've opened bug 1573600 for that issue
Tal, I've ran the scenario in comment 0 with build rhvm-4.2.3.5-0.1.el7.noarch. Now, there is no NPE, and the VM is registered, but the engine log shows that it fails to add the memory snapshots, this patch only takes care of the vm's registration? (not including the memory snapshots?)
Yes, the importing of memory snapshot was broken from day one, that's why I've opened bug 1573600 for
Following comment 8 and comment 9 moving to verified.
This bugzilla is included in oVirt 4.2.3 release, published on May 4th 2018. Since the problem described in this bug report should be resolved in oVirt 4.2.3 release, it has been closed with a resolution of CURRENT RELEASE. If the solution does not work for you, please open a new bug report.