ESXi Import Wizard: create storage failed: failed to spawn fuse mount...?

[ronejamesdo]

I have two ESXi 6.7U3 boxes. One works fine with Datacenter -> Storage -> Add -> ESXi, the other results in this:



Does anybody know how to fix this?

This got some mentions (#34, #155, #230) in the Import Wizard announcement thread (https://forum.proxmox.com/threads/n...ing-vmware-esxi-based-virtual-machines.144023) but I am not sure if it was answered, that thread has a lot in it at this point.

Thank you.

RJD

 

[dcsapak]

does the journal during that time say anything?

 

[ronejamesdo]

does the journal during that time say anything?

Presuming there's not some different log or journal for Proxmox which you are referencing, not much (journalctl -f):

May 02 06:26:49 proxmoxserver01 systemd[1]: Started pve-esxi-fuse-esxiserver001.scope.
May 02 06:26:50 proxmoxserver01 systemd[1]: pve-esxi-fuse-esxiserver001.scope: Deactivated successfully.
May 02 06:26:50 proxmoxserver01 systemd[1]: run-pve-import-esxi-esxiserver001-mnt.mount: Deactivated successfully.

 

[dcsapak]

are you sure that's all? can you post some time before & after those log lines too?

 

[ronejamesdo]

100% sure. Everything around it is repeated over and over through the journal. Most of it relates to logins for a monitoring software that is watching the box. There is no difference in what happens before and after those three lines. (The monitor checks every minute, if I time it to about right about the 30-second mark then the above logs have a 30 second lull before and after.)

 

[dcsapak]

mhm seems we don't have enough logging at the point where it fails in your setup

can you post your network/etc. config of the pve and esxi ? is there any firewall in between? anything special on the esxi side?

 

[ronejamesdo]

Nothing that I know of is separating the two machines (which are on the same subnet, a couple of IPs away from one another and physically close as well), I can ssh to the ESXi from the Proxmox for example. I am not sure if you are looking for a screenshot, contents of a specific file, or output of a command with regard to posting the network/etc config.

I understand that Proxmox leverages ESXi's API to connect and migrate. Do you happen to know if there's a "whether the API is working rightly" test that I can do (or is logging into ESXi via a browser test enough)?

 

[ronejamesdo]

A recent poster in the import wizard announcement thread with the same thing said that it ended up having to with "due to having the same name for 2 different type of datastore on the ESXi" which I assume means under "Storage" and "Datastores" in his ESXi. My ESXi has only:

SSD Raid6 + 2 spares

...but I wonder if you'd know if it could have to do with the " + " maybe (like that's a special character that won't be auto escaped)?

Is there somewhere else in Proxmox that might log something like that?

 

[ronejamesdo]

I confirmed this today. Renaming its datastore to "SSD Raid6 and 2 spares" (eliminating the "+") fixed it for me (and took only a second and only required the ESXi GUI).

 

[dcsapak]

well great that you figured out what the problem was, i'll see if we can fix that in our code (if it is indeed our code) and maybe add a bit more verbose output to the logs so that we can see such things better in the future

 

[dcsapak]

ok i tested a bit and it seems that the esxi http api we use here cannot properly handle the '+' symbols (others are not a problem e.g. space or '\')
not sure if we can do anything from our side here

i sent two patches to improve the error logging/printing so that we maybe catch such error easier in the future:
https://lists.proxmox.com/pipermail/pve-devel/2024-May/063812.html

 

[LeonidR]

ProxMox 8.2.4
VMWare VSphere 8.0.3 HA Cluster on two ESXi nodes

A similar situation with the same error as in the picture in the first message when trying to connect one node of the ESXi from cluster.

At the moment of connection to ESXi cluster node on PpoxMox GUI we see an error message and in the the logs
#journalctl -f :

Sep 07 11:02:17 pve systemd[1]: Started pve-esxi-fuse-ESXI82.scope.
Sep 07 11:02:17 pve esxi-folder-fuse[251900]: Error: error when getting file size: "ha-datacenter"/""/"/var/run/crx/infra/vCLS-cae84d56-bdf0-6af6-748e-2ac9abc811a6/vCLS-cae84d56-bdf0-6af6-748e-2ac9abc811a6.vmx" file not found
Sep 07 11:02:17 pve systemd[1]: pve-esxi-fuse-ESXI82.scope: Deactivated successfully.
Sep 07 11:02:17 pve systemd[1]: run-pve-import-esxi-ESXI82-mnt.mount: Deactivated successfully.

ProxMox can't determine size\datastore vCLS VM in cluster