It seems to be the same problem.
Now the question is, which numbers to trust? Does any of them show the correct value how much is actual used and how much is actual still free?
For the free space this works for docker nodes:
docker exec -it storagenode /app/dashboard.sh
The GUI dashboard shows for me completely nonsense values.
Theoretically the multinode dashboard should be correct.
Did it match values on the multinode dashboard?
But 1.64TB used + 6.25GB free + 65.47GB trash != 2TB
So from the math, the node dashboard is more accurate. Where did the space go that is missing from the multinode dashboard?
This is why I believe that there is a bug anyway.
How a CLI dashboard shows it?
CLI shows 300GB lower available space than the node dashboard, e.g. 450GB vs 750GB free.
I don’t use the multinode dashboard, but from my experience the CLI dashboard shows ~5GB free on a full node (ingress has stopped) while the GUI dashboard shows still free space.
Storage Node Dashboard ( Node Version: v1.137.5 )
======================
ID 1ydNTgYoCUTxB31dXrCXxu4tPX4T4phJaWRztfQmdvZq6CPXrY
Status ONLINE
Uptime 431h51m56s
Available Used Egress Ingress
Bandwidth N/A 180.91 GB 109.46 GB 71.45 GB (since Oct 1)
Disk 4.98 GB 1.04 TB
Here the GUI:
From a bigger Node:
Storage Node Dashboard ( Node Version: v1.137.5 )
======================
ID 1gr19eA1Sx1gkNhWUW3WjULhaTgvLnXoCjJXinGNXA8y7xosA3
Status ONLINE
Uptime 170h18m8s
Available Used Egress Ingress
Bandwidth N/A 0.85 TB 539.76 GB 314.16 GB (since Oct 1)
Disk 269.72 GB 5.95 TB
Here the GUI:
Used, Free and Trash sum up to 7.4TB, so the calculation is fine.
My Nodes are all migrated to Hashstore and i think the difference between CLI and GUI for the free space lies in the not compacted LOG files.
Hashstore stats:
2025-10-13
Node ID: 1gr19eA1Sx1gkNhWUW3WjULhaTgvLnXoCjJXinGNXA8y7xosA3
Hashstore: AP1
238981732 NodePath/storage/hashstore/SatAP1/
Files:
257
Hashstore: EU1
1703842740 NodePath/storage/hashstore/SatEU1/
Files:
1673
Hashstore: US1
5016288664 NodePath/storage/hashstore/SatUS1/
Files:
4865
This sums up to 6,959,113,136 Byte or ~6.959TB physically used space by Hashstore.
Perhaps we need an extra category for the not yet recovered free space in the LOG files?
So, seems the CLI and multinode dashboards are correct about the free space. An you are correct, seems the not cleaned space is just occupy the disk space, while it’s considered as “free”.
Maybe it should be considered the same as trash? Occupied but to be freed.
But does it also mean we have a 4 tier step now how space finally gets freed? 1st data that is to be removed gets marked for removal, then 2nd garbage collected via bloomfilter, then it remains in trash for 7 days and will only really get removed from disk after log compaction? Does it mean from the moment the customer deletes a file to final deletion on the node we are talking about multiple weeks now?
Likely more than a week for sure. But the compaction frequency and the percentage are configurable