Is 10x Compression Normal?

thomasjcf21

Dabbler
Joined
Jun 19, 2019
Messages
15
Hi Guys,

So yesterday I moved all my VMs from a rubbish SAN over to FreeNAS, however, I've noticed that ESXi is showing that I've used 5TB of data where FreeNAS is showing 500GB data used.

I know that compression is enabled on the Zpool however, I didn't expect it to be as good as 10x compression rate. De-dup is not enabled on the pool as far as I'm aware.

I don't know if there's an issue with FreeNAS and data might be lost or if this is expected? I've attached screenshots below.

They show the overall datastore cluster / datastore usage as well as what FreeNAS is showing.

Many Thanks,

Tom.
 

Attachments

  • Screenshot 2019-06-27 at 10.58.33.png
    Screenshot 2019-06-27 at 10.58.33.png
    197.3 KB · Views: 378
  • Screenshot 2019-06-27 at 10.58.49.png
    Screenshot 2019-06-27 at 10.58.49.png
    256.9 KB · Views: 379
  • Screenshot 2019-06-27 at 10.58.57.png
    Screenshot 2019-06-27 at 10.58.57.png
    127.1 KB · Views: 360
  • Screenshot 2019-06-27 at 10.59.04.png
    Screenshot 2019-06-27 at 10.59.04.png
    54.1 KB · Views: 371
  • Screenshot 2019-06-27 at 10.59.12.png
    Screenshot 2019-06-27 at 10.59.12.png
    99.9 KB · Views: 369
  • Screenshot 2019-06-27 at 10.59.18.png
    Screenshot 2019-06-27 at 10.59.18.png
    146.9 KB · Views: 365

rvassar

Guru
Joined
May 2, 2018
Messages
972
Thin provisioning maybe?
 

thomasjcf21

Dabbler
Joined
Jun 19, 2019
Messages
15
Thin provisioning maybe?

The zvols presented to ESXi are thin provisioned and some of the VMs are thin provisioned. However, there's a 2TB backup VM that's been transferred across and that's not really showing an impact.
 

rvassar

Guru
Joined
May 2, 2018
Messages
972
That is odd. I think I'd verify each vmdk in the datastore browser for starters, but I'm somewhat stumped. Is there a discrepancy between the free space reported in the GUI and the CLI tools? Both zpool and even a simple "df -h"?
 

thomasjcf21

Dabbler
Joined
Jun 19, 2019
Messages
15
That is odd. I think I'd verify each vmdk in the datastore browser for starters, but I'm somewhat stumped. Is there a discrepancy between the free space reported in the GUI and the CLI tools? Both zpool and even a simple "df -h"?

All the VMDK's look to be ok, I've turned on random VMs and all seems to be working. As the backup server isn't my data I'm not able to access it to verify it. But it appears to also have been successful.

I don't get how there's 10x compression, to me that doesn't make sense!!!! :')
 
Top