Space usage discrepancy between cluster filer and Linux client

[heightsnj]

vol show xyz -field used --> 8.61TB, and total size assigned -->50TB ( NFS Volume)

However, on the linux client. df -h shows 24TB used. Size Used Avail Use% Mounted on 49T 24T 26T 48% /mount/xyz

This volume has "auto" tiering, and deduplication enabled. I believe the figure on the filer is accurate.

1. If it is, then why did Linux client show such large amount of difference?

2. how and what commands can tell me the accurate usage on Linux?

Thank you!

[paul_stejskal]

What is output of "volume show-footprint" and "volume show-space"?

[heightsnj]

Please find the following:

:*> volume show-footprint xyz

Vserver : svm1
Volume : xyz

Feature Used Used%
-------------------------------- ---------- -----
Volume Data Footprint 8.03TB 25%
Footprint in Performance Tier
6.20TB 74%
Footprint in S3Bucket 2.15TB 26%
Volume Guarantee 0B 0%
Flexible Volume Metadata 244.4GB 1%
Deduplication 39.25GB 0%
Delayed Frees 326.8GB 1%

Total Footprint 8.62TB 27%

*> volume show-space xyz

Vserver : svm1
Volume : xyz

Feature Used Used%
-------------------------------- ---------- ------
User Data 7.81TB 16%
Filesystem Metadata 39.03GB 0%
Inodes 21.39MB 0%
Snapshot Reserve 1TB 2%
Deduplication 174.9GB 0%
Performance Metadata 6.46GB 0%

Total Used 9.03TB 18%

Total Physical Used 8.03TB 16%

[paul_stejskal]

Odd. I'd open a case as we need a packet trace to see why it is doing that.

[heightsnj]

If I told you that this is in a cloud environment, the filer is CVO, and the Linux server is a cloud VM. Would that make any different to you?

I thought it wouldn't, that's why I didn't mention it earlier. 

[paul_stejskal]

No it wouldn't.