Disk Error - VM Pauses and cant be recovered

[Al_AU]

Hi All,
I am running a virtual device on Ubuntu 20.04 installed with Virtualbox 6.1 and the extention pack.

I have set the guest up with a second (DATA) hard disk of 17TB Ext4 on a 25TB disk on the host server. This is Isolated from the Guest OS

on the systems I have tested I am getting the same result.
The Guest data disk (created as 17TB VDI) is showing as 17.6TB file on disk.
In the Virtual box media manager the disk shows as 17TB and 16TB in use.
The host shows 5.1TB free on the Host drive (there are a couple of other files on there.)

On Restart the guest device boots and allows me to log in but immediately throws an error and pauses the device
An error has occurred during virtual machine execution! (rest of the base message)
The I/O cache encountered an error while updating data in medium "lsilogicscsi-0-1"
(rc=VERR_FILE_TOO_BIG). Make Sure there is enough free space on the disk and that the disk is working properly. Operation can be resumed afterwards.

Error ID: BLKCACHE_IOERROR
Severity: Non-Fatal Error

The guest runs well for weeks then suddenly hits this wall. Im at a loss because there seems to be more than sufficient free space and I have recreated the guest disk 3 times, each time 1TB smaller to rule out the host disk reaching capacity.

[mpack]

17TB? That is a really huge virtual hard drive! I don't have any PC drives large enough to hold a 17TB file. Even my NAS is only 21TB usable, about 13TB unused.

Why so large? It seems bound to cause problems. In this case if you have a dynamic VDI and the host free space is less than X percent of the file then I would not be surprised to get warnings.

Having a bulk storage drive is one thing. Dedicating it to a single VM creates many problems, like how on earth do you intend to back up this VM? There's a phrase involving eggs and baskets that seems like it might be appropriate here...

[fth0]

You could try the following and report back if it helps:

In the VM configuration, enable Use Host I/O Cache for the LsiLogic Controller.

[mpack]

I somehow doubt that VERR_FILE_TOO_BIG will be fixed by changing a cacheing option.

Ext4 usually has a max file size of 16TB, the OP says he configured a 17TB VDI. I expect that the problem will be in there somewhere, though at this moment it looks like Ext4 is the guest rather than host filesystem.

[fth0]

The background reason for my suggestion was that the error message was reported by the BlockCache implementation. But like you, I assume that the error originally came from the host OS (as errno == 27 == EFBIG).

[Al_AU]

Thanks for your help all,
fth0, I did give it a try and it made no difference to the operation unfortunately

The answer came after more forum searching and it is exactly as mpac posted (no not the bit where it sounded like my GF "That is a really huge..." )

Ext4 usually has a max file size of 16TB.

The virtual appliance (an NVR) apparently will find and configure all attached drives on install so I created 3 x 7TB drives and let the appliance do its thing.

Thanks again, legendary help