Home > Storage > PowerScale (Isilon) > Product Documentation > Data Protection > Data Protection with Dell PowerScale SnapshotIQ > Writing to a snapshot
This following figure illustrates the process of updating a file’s metadata, by way of a simple UID change. When this occurs, and OneFS detects that it needs to duplicate the old data, it copies the old blocks to the snapshot. In this case, there was no LIN 1:abcd:1234/98 prior to the write, so any attempt to read this LIN would have fallen forward to the head version.
When a file’s metadata is changed, OneFS needs to create a copy of the data, and increment the Snap_ID. Since there is no actual data written, there is no need to copy over any old data blocks to the new snapshot, so ditto-blocks are used instead. However, the first time a file is changed after a snapshot is taken, the LIN is recorded in the snapshot tracking file (STF). This STF lets OneFS efficiently know what data can be removed when a snapshot is deleted.
The file being written to is unchanged since the last snapshot. The data blocks from the head version of the file are copied to the snapshot version, replacing the ditto-blocks, and the head version is overwritten with the new data. This is an example of copy on write (COW).