0

I am confused about what Snapshot Copy Reserve actually stores. From what I understand, a snapshot copy is a copy of pointers which points to blocks on disk which were occupied by a file at some point in time. File blocks are stored on volume and pointers are stored on Snapshot Copy Reserve. When a file is modified, a new version of data from a modified block is written to another block (in volume). The original block stays in read-only mode for as long as the first snapshot exists. Modified block is also written to volume and the snapshot of the modified file is written to the Snapshot Copy Reserve.

Now, it doesn't make sense to me how and why snapshot spills happen. Shouldn't volume run out of space before snapshot spill in all cases?

Alex
  • 153
  • 5

1 Answers1

0

Assume for a moment, a 1TB volume with a 10% Snapshot Copy Reserve, exposed as an NFS export.

The usable space on the volume is now 900GB. The snapshot reserve is 100GB. If you were to view the volume on an attached server, you'd see it with 900GB of free space.

What "snapshot spill" is, in NetApp land, is when the size of the snapshot runs over the Snapshot Copy Reserve. So if you snapshot the volume, then write/change 120GB of data to the volume, the snapshot will hit 120GB of space used. The Snapshot Copy Reserve does NOT prevent a snapshot from using more space than is allocated to the reserve. In this instance, if you again looked at the volume from a mounted server, you'd see it with 880GB free.

Driftpeasant
  • 3,217
  • 2
  • 22
  • 28