New Import Wizard Available for Migrating VMware ESXi Based Virtual Machines

[lknite]

Clean install of the latest proxmox, configured network w/ home and storage network (used linux bridges); able to mount an nfs share via the storage network (which confirms i have the needed network access), added storage type of 'esxi' for my two esxi hosts running esxi 8u3, clicked on each of the esxi hosts and i can see the vms running there; used esxi to shutdown a small basic vm, went back to proxmox, did a shift-f5 to reload the page, clicked on the vm, clicked Import and I get:
```storage 'esxi-two' is not activated (500)```

the vm is running linux with docker and a single container; the storage is local storage on the esxi on an nvme; esxi hosts have a valid license; the vm has no snapshots

 

[Falk R.]

Clean install of the latest proxmox, configured network w/ home and storage network (used linux bridges); able to mount an nfs share via the storage network (which confirms i have the needed network access), added storage type of 'esxi' for my two esxi hosts running esxi 8u3, clicked on each of the esxi hosts and i can see the vms running there; used esxi to shutdown a small basic vm, went back to proxmox, did a shift-f5 to reload the page, clicked on the vm, clicked Import and I get:
```storage 'esxi-two' is not activated (500)```

the vm is running linux with docker and a single container; the storage is local storage on the esxi on an nvme; esxi hosts have a valid license; the vm has no snapshots

Is the ESXi enabled in the Storage tab under DataCenter?

 

[lknite]

Is the ESXi enabled in the Storage tab under DataCenter?

yes

 

[lknite]

Is the ESXi enabled in the Storage tab under DataCenter?

maybe this is useful?
Sep 27 14:23:31 pve-c esxi-folder-fuse[25953]: Error: error when getting file size: "ha-datacenter"/""/"/var/run/crx/infra/vCLS-5159a1a8-3af3-0000-0000-000000000000/vCLS-5159a1a8-3af3-0000-0000-000000000000.vmx"
file not found
Sep 27 14:23:31 pve-c systemd[1]: pve-esxi-fuse-esxi\x2done.scope: Deactivated successfully.

I saw 'ha' so I enabled ha in vcenter and tried again, still failed, i just disabled it again ... it wasn't running ha before

 

[Falk R.]

maybe this is useful?
Sep 27 14:23:31 pve-c esxi-folder-fuse[25953]: Error: error when getting file size: "ha-datacenter"/""/"/var/run/crx/infra/vCLS-5159a1a8-3af3-0000-0000-000000000000/vCLS-5159a1a8-3af3-0000-0000-000000000000.vmx"
file not found
Sep 27 14:23:31 pve-c systemd[1]: pve-esxi-fuse-esxi\x2done.scope: Deactivated successfully.

I saw 'ha' so I enabled ha in vcenter and tried again, still failed, i just disabled it again ... it wasn't running ha before

You should not touch the VCLS VMs.

If your vSphere cluster does not use DRS, simply have the VCLS VMs removed.
Navigate to the vCenter Server and then toConfiguretab.
Click onAdvanced settingsection and then onEdit settingsbutton.
Add a new entry with name = config.vcls.clusters.domain-c<number>.enabled and value = False.
You found the Domain Number in your Browser URL, when you select the Cluster.

 

[lknite]

You should not touch the VCLS VMs.

If your vSphere cluster does not use DRS, simply have the VCLS VMs removed.
Navigate to the vCenter Server and then toConfiguretab.
Click onAdvanced settingsection and then onEdit settingsbutton.
Add a new entry with name = config.vcls.clusters.domain-c<number>.enabled and value = False.
You found the Domain Number in your Browser URL, when you select the Cluster.

I do currently have DRS enabled, i'll try turning it off

 

[sw-omit]

I do currently have DRS enabled, i'll try turning it off

On top of turning it off (or at least putting it to "manual" or "only on startup moving") to prevent sudden movement of VM's, you don't have to migrate over the vCLS VM's at all, as they have no use on Proxmox. Those VM's are only managed by vCenter to help with DRS transfers, they are not VM's you set up yourself (if you remove them manually from vmware/vcenter, they will be automatically re-created.)

 

[lknite]

I wasn't trying to move the vCLS vm ... was trying to move a basic linux vm as described before ... posted some logs hoping to troubleshoot, maybe the import gui was trying to move vCLS vm for some reason ... in any case, i gave up and have been exporting the vms one at a time and importing them in proxmox with 'qm import ovf ...' ... i've only got about 20, this will work.

 

[Falk R.]

I wasn't trying to move the vCLS vm ... was trying to move a basic linux vm as described before ... posted some logs hoping to troubleshoot, maybe the import gui was trying to move vCLS vm for some reason ... in any case, i gave up and have been exporting the vms one at a time and importing them in proxmox with 'qm import ovf ...' ... i've only got about 20, this will work.

You are required to have the VCLS VMs deleted from vCenter, which works with the above instructions.
Deactivating DRS is not sufficient, as the VCLS VMs continue to run in the background.
The Proxmox Migration Tool tries to read in all .vmx files, which does not work with the VCLS machines.

 

[sw-omit]

You are required to have the VCLS VMs deleted from vCenter, which works with the above instructions.
Deactivating DRS is not sufficient, as the VCLS VMs continue to run in the background.
The Proxmox Migration Tool tries to read in all .vmx files, which does not work with the VCLS machines.

That is strange, because for me that never was an issue during migration.
I had DRS set to manual, but the vcls-VM's still there, they even showed up in the list for VM's to be able to be migrated (because one of them was on the ESXi-server I connected to), I just of course didn't select them to migrate because of the reasons above, but I did notice that disk-name in the error, which is why I added my note.

 

[Falk R.]

That is strange, because for me that never was an issue during migration.
I had DRS set to manual, but the vcls-VM's still there, they even showed up in the list for VM's to be able to be migrated (because one of them was on the ESXi-server I connected to), I just of course didn't select them to migrate because of the reasons above, but I did notice that disk-name in the error, which is why I added my note.

Of course, this can also work with the VCLS, especially if you integrate the ESXi directly in the PVE.
The vCenter always wants to hide the VCLS somewhat, which can cause problems.

 

[shep]

Related issue:

https://forum.proxmox.com/threads/error-when-accessing-vmware-storage.155331/

 

[patwes]

Hi
Anyone having problems specifically esxi 8.0U3 ? upgraded from U2 where import worked, now i get

storage 'pesx04' is not activated (500) (and it is, and i can list vm:s)

get this every time i try to import a vm, have removed connection and re added, also rebooted host, both vmware and proxmox, all are running latest releases as today
anyone else tried it om U3b ?

have verified that it works on 8U1 and 8U2

Regards Patrik

 

[Falk R.]

I'm afraid this is due to the API changes that were introduced with 8.0U3.
The importer was built based on the 8.0U2 API documentation. I hope that this problem is already known and there will be an update for the importer soon.

 

[SpyderDK]

Has there been any update on the issue regarding esxi 8.0U3 and the importer?

 

[sie]

Import fails for me before it even really starts (in seconds)

vSphere ESXi-8.0.0-20513097-standard Host (Host was rebooted after making configuration changes)
Config.HostAgent.vmacore.soap.maxSessionCount = 0
Config.HostAgent.vmacore.soap.sessionTimeout = 0

Proxmox VE 8.2.2

Formatting '/var/lib/vz/images/123/vm-123-disk-0.qcow2', fmt=qcow2 cluster_size=65536 extended_l2=off preallocation=off compression_type=zlib size=131072 lazy_refcounts=off refcount_bits=16
transferred 0.0 B of 128.0 KiB (0.00%)
transferred 128.0 KiB of 128.0 KiB (100.00%)
transferred 128.0 KiB of 128.0 KiB (100.00%)
efidisk0: successfully created disk 'local:123/vm-123-disk-0.qcow2,size=128K'
create full clone of drive (esxinode1:ha-datacenter/node1-datastore2/Test/Test.vmdk)
Formatting '/var/lib/vz/images/123/vm-123-disk-1.qcow2', fmt=qcow2 cluster_size=65536 extended_l2=off preallocation=off compression_type=zlib size=1536 lazy_refcounts=off refcount_bits=16
transferred 0.0 B of 1.5 KiB (0.00%)
qemu-img: output file is smaller than input file
TASK ERROR: unable to create VM 123 - cannot import from 'esxinode1:ha-datacenter/node1-datastore2/Test/Test.vmdk' - copy failed: command '/usr/bin/qemu-img convert -p -n -f vmdk -O qcow2 /run/pve/import/esxi/esxinode1/mnt/ha-datacenter/node1-datastore2/Test/Test.vmdk zeroinit:/var/lib/vz/images/123/vm-123-disk-1.qcow2' failed: exit code 1

Same issue occurs with vCenter as Storage instead. I'm not doing a Live Migration, VM is off on the vSphere ESXi host. VM has a EFI disk because it utilises UEFI, not using a TPM though. VM has no snapshots. The VM lives on a local datastore on the ESXi host, I've tried transferring it to a local Proxmox datastore and a ZFS one.

Could this be a networking issue? An inability to contact the appropriate services on the relevant VMkernel NIC of the ESXi Host? Could any such requirements be documented at https://pve.proxmox.com/wiki/Migrate_to_Proxmox_VE please.

 

[Falk R.]

Import fails for me before it even really starts (in seconds)

vSphere ESXi-8.0.0-20513097-standard Host (Host was rebooted after making configuration changes)
Config.HostAgent.vmacore.soap.maxSessionCount = 0
Config.HostAgent.vmacore.soap.sessionTimeout = 0

Proxmox VE 8.2.2

Formatting '/var/lib/vz/images/123/vm-123-disk-0.qcow2', fmt=qcow2 cluster_size=65536 extended_l2=off preallocation=off compression_type=zlib size=131072 lazy_refcounts=off refcount_bits=16
transferred 0.0 B of 128.0 KiB (0.00%)
transferred 128.0 KiB of 128.0 KiB (100.00%)
transferred 128.0 KiB of 128.0 KiB (100.00%)
efidisk0: successfully created disk 'local:123/vm-123-disk-0.qcow2,size=128K'
create full clone of drive (esxinode1:ha-datacenter/node1-datastore2/Test/Test.vmdk)
Formatting '/var/lib/vz/images/123/vm-123-disk-1.qcow2', fmt=qcow2 cluster_size=65536 extended_l2=off preallocation=off compression_type=zlib size=1536 lazy_refcounts=off refcount_bits=16
transferred 0.0 B of 1.5 KiB (0.00%)
qemu-img: output file is smaller than input file
TASK ERROR: unable to create VM 123 - cannot import from 'esxinode1:ha-datacenter/node1-datastore2/Test/Test.vmdk' - copy failed: command '/usr/bin/qemu-img convert -p -n -f vmdk -O qcow2 /run/pve/import/esxi/esxinode1/mnt/ha-datacenter/node1-datastore2/Test/Test.vmdk zeroinit:/var/lib/vz/images/123/vm-123-disk-1.qcow2' failed: exit code 1

Same issue occurs with vCenter as Storage instead. I'm not doing a Live Migration, VM is off on the vSphere ESXi host. VM has a EFI disk because it utilises UEFI, not using a TPM though. VM has no snapshots. The VM lives on a local datastore on the ESXi host, I've tried transferring it to a local Proxmox datastore and a ZFS one.

Could this be a networking issue? An inability to contact the appropriate services on the relevant VMkernel NIC of the ESXi Host? Could any such requirements be documented at https://pve.proxmox.com/wiki/Migrate_to_Proxmox_VE please.

Hi, your error message is related to the qcow2 image. If you use ZF, then import into the ZFS pool instead of a dataset and leave the disk in RAW format.

 

[sie]

I had tried transferring it to a local non-ZFS Proxmox datastore with the same result.

 

[Falk R.]

I had tried transferring it to a local non-ZFS Proxmox datastore with the same result.

Also as qcow2 Image? Any Errormessage?

It looks like an error I have seen before. Sometimes VMs have already been converted to vSphere before and the VMDK has a size with a few digits after the dot. Like 100.107356736 GB. This always causes problems when converting. I enlarge these disks under vSphere to whole GB, e.g. 101 GB. Then it works better.
I also always migrate to diskpools (ZFS, Ceph RBD or LVM) and do not use qcow2. Therefore, I have never had such an error message.

 

[sie]

I have discovered the problem. The VM had a VM Storage Policy applying to it, the VM was encrypted using the VMware VM Encryption Provider. Removing the policy, and thus the encryption, allowed it to be imported. The preparation section of the wiki page should include checking for such policies.