#20008 new defect
VERR_SSM_LOAD_CONFIG_MISMATCH on vagrant snapshot restore
| 回報者: | Marc Günther | 負責人: | |
|---|---|---|---|
| 元件: | VM control | 版本: | VirtualBox 6.1.16 |
| 關鍵字: | snapshot, vagrant | 副本: | |
| Guest type: | Linux | Host type: | Mac OS X |
描述
When restoring a VM snapshot using vagrant snapshot restore .., I get the following error:
executing: vagrant snapshot restore 4-st_6_0_0_20-9f29a314b9e1f5626f0f27ebfa8fc0a92dd66e42 ==> default: Forcing shutdown of VM... ==> default: Restoring the snapshot '4-st_6_0_0_20-9f29a314b9e1f5626f0f27ebfa8fc0a92dd66e42'... Progress: 0% Progress: 90% ==> default: Resuming suspended VM... ==> default: Booting VM... err: There was an error while executing `VBoxManage`, a CLI used by Vagrant for controlling VirtualBox. The command and stderr is shown below. Command: ["startvm", "4768f166-7914-4c98-9e81-c4f351614e51", "--type", "headless"] Stderr: VBoxManage: error: lsilogicscsi#0: Target 0 config mismatch: config=false state=true [ver=5 pass=final] (VERR_SSM_LOAD_CONFIG_MISMATCH) VBoxManage: error: Details: code NS_ERROR_FAILURE (0x80004005), component ConsoleWrap, interface IConsole
When I try the displayed startvm command directly, I get the same error (I attached the complete log of that run to this ticket):
$ VBoxManage startvm "4768f166-7914-4c98-9e81-c4f351614e51" --type headless Waiting for VM "4768f166-7914-4c98-9e81-c4f351614e51" to power on... VBoxManage: error: lsilogicscsi#0: Target 0 config mismatch: config=false state=true [ver=5 pass=final] (VERR_SSM_LOAD_CONFIG_MISMATCH) VBoxManage: error: Details: code NS_ERROR_FAILURE (0x80004005), component ConsoleWrap, interface IConsole $
Same thing happens in the GUI, when I start the VM or Restore the snapshot.
BUT, when I Discard the saved state, the VM starts up normally. Of course, this doesn't help me, as I cannot use the GUI (see below).
When I try to Restore again, it happens again.
This is VirtualBox 6.1.16, Vagrant 2.2.10 on MacOSX 10.12. But it has happened in this form since at least VirtualBox 6.0.x.
Background Info:
We have a script which automatically provisions a VM to contain our applications, database and everything we need to run our E2E-Tests against.
In order to speed up the provisioning, the script is split into several steps. As some steps can be quite slow, each step saves a VM snapshot at the end, so on the next run of the script, if nothing has changed, it can simply restore the snapshot instead of re-running that step again. (similar idea to the lines in a Dockerfile)
The above error mostly happens on the second run of the script, when it wants to restore the snapshot it has just taken from the last step on the first run.
The problem seems to be with that particular snapshot. When I delete it, the script will restore a previous snapshot, which works fine. It will then run the last step again, and snapshot it, and that snapshot usually is also fine.
This is happening for several months now. An interesting pattern I noticed, this only seems to happen when the last step is a no-op (which means there were two snapshots created in quick succession). It is then always this second snapshot that is broken. The previous snapshot is fine. Also when the last step is not a no-op, the error will not occur.
There is nothing peculiar about taking the snapshots, I simply issue vagrant snapshot save default <snapshot_name> each time.
log file of: VBoxManage startvm "4768f166-7914-4c98-9e81-c4f351614e51" --type headless