Description of problem: A disk move between storage domains has resulted in the image being removed from the source SD before doing the copy/convert(qemu-img) operation to the destination host. Version-Release number of selected component (if applicable): ovirt-engine-4.1.8.2-0.1.el7.noarch vdsm-4.19.43-3.el7ev.x86_64 FC Block storage How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: - The disk image was removed from source for which VM is missing disk and unable to boot. Expected results: - The disk image to be copied to the destination SD and then remove from source. Additional info:
From the log it seems that the move command was run twice with the same parameters. It is not possible to perform with the REST API. Though the user mentioned he might have click on the UI dialogue twice. I will try to reproduce on 4.1.
I can reproduce in 4.1. Though using the debugger for update the command parameter of the source storage domain, not possible to do with REST API. The move operation was performed twice with the same parameters from the UI. I don't know how it happened, but the reproduction shows the exact same flow in the logs. I will try to understand now how this flow affected the volume.
This is the flow of this is issue: User move disk A from SD_src to SD_dst via UI. Move succeeded. Disk A is now on SD_dst, the disk on SD_src is deleted. Again,(by mistake/UI issue) User move disk A from SD_src to SD_dst via UI. The engine tries to create a volum on SD_dst but fails, because it is already existing there. Then the MoveImageGroupCommand execute endWithFailure that deletes the disk in SD_dst which was the current disk copy. ( I guess trying to cleanup the failure). Now the disk is gone... Now, when the VM tries to start, it fails to find the volume, because it was deleted. I don't know what issue happened in the UI. The user mentioned in the case: "I might have click twice on OK button of the move dialogue, since it didn't disappear when I click the first time." The one thing I can think of is to add some validation in the Move command to avoid thi flow again. Note that in 4.2, the command fails before trying to create the destination disk so it is cannot be reproduce there.
In 4.2 actually , it will also delete the disk on failure. Should be solved also there.
*** Bug 1540351 has been marked as a duplicate of this bug. ***
INFO: Bug status (ON_QA) wasn't changed but the folowing should be fixed: [Project 'ovirt-engine'/Component 'vdsm' mismatch] For more info please contact: rhv-devops
Verified with the following code: --------------------------------------- ovirt-engine-4.2.2.1-0.1.el7.noarch vdsm-4.20.19-1.el7ev.x86_64 Verified with the following scenario: --------------------------------------- 1. Create a VM with disks 2. Move disk to another domain >>>>> When pressing OK the Move Disk dialogue closes immediately and does not allow pressing OK twice. Moving to VERIFIED
INFO: Bug status (VERIFIED) wasn't changed but the folowing should be fixed: [Project 'ovirt-engine'/Component 'vdsm' mismatch] For more info please contact: rhv-devops
*** Bug 1574346 has been marked as a duplicate of this bug. ***
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHEA-2018:1489
BZ<2>Jira Resync
sync2jira