--- title: "What happens when you scan a drive · Space Wizard" description: "The walk, the totals and the duplicate check, locally and on a remote host." url: "https://royalsoftworks.com/products/space-wizard/how-it-works/the-scan/" source: "https://royalsoftworks.com" format: "markdown" note: "Markdown rendering of the HTML page at `url`. Same content, same canonical URL." --- How it works · 1 of 2 # What happens when you scan a drive. The walk, the totals and the duplicate check, locally and on a remote host. From click to cleanup ## What happens when you scan a drive. 1. 01 ### Pick a source A local drive, or a saved remote host. Free scans your home drive; Pro opens every local drive on the machine. Remote hosts work on either tier. 2. 02 ### Cloud-sync folders are skipped OneDrive, Dropbox, iCloud, Google Drive, Box, pCloud and MEGA are excluded by default — their placeholder files aren't real local bytes, and walking them can make the provider download the real thing. One switch includes them if you want. 3. 03 ### The walk runs, 64 folders at a time Locally, a concurrent breadth-first walk reads up to 64 directories at once and reports each folder's direct size as it goes, so the running total climbs while the scan is still working. Junctions and symlinks are resolved so nothing is counted twice. 4. 04 ### Nothing is quietly dropped Files the kernel refuses to open — pagefile.sys and friends — are sized from the directory entry instead, roughly 15 GB most scanners lose on a system drive. The app asks for Administrator rights at launch; folders that still couldn't be read are counted and surfaced. 5. 05 ### Remote hosts take the same path, differently On a host with a shell, a single \`find\` command streams type, size and path for the whole tree in one pass. Without one — a locked-down SFTP account, or plain FTP — it falls back to a directory walk over as many parallel connections as that host allows. 6. 06 ### Duplicates get a three-pass check Files are grouped by exact byte size; unique sizes are dropped immediately. Survivors get a SHA-1 of their first 4 KB. Only what still matches is hashed in full. Over SSH, both hash passes run on the host, so hashes cross the wire rather than files; FTP and shell-less accounts read the candidates instead. 7. 07 ### You see bars or a donut The same result renders as a bar chart or a donut of the folder you're standing in — the nine largest children plus an "Other" bucket. Open one and the chart follows you down. 8. 08 ### Clear what you don't need Multi-select in the Duplicates tab or browse the Files tab directly. At least one copy in every duplicate group is always kept. Local deletions go to the recycle bin; remote deletions are permanent, and the button says so. For your IT team - **Local walk** — Node.js concurrent breadth-first walk, 64-way, streaming per-directory results in batches; depth capped at 512 - **Remote walk** — GNU \`find -printf\` single stream where available; otherwise a concurrent readdir walk over 1–16 connections (default 5) - **Duplicate hash** — Exact byte size → SHA-1 over the first 4,096 bytes → full SHA-1 on survivors, 8 files at a time, largest first - **Remote hashing** — \`dd bs=4096 count=1 | sha1sum\` for the prefix pass and \`sha1sum\` for the full pass, executed on the remote host - **Deletion** — OS trash / recycle-bin API locally; direct remove over SFTP or FTP on a remote host - **Licence gate** — Free: home drive only, 100 GB measured per scan. Pro: every local drive, no cap. Remote hosts on both. [Previous How it works: all 2 pages](https://royalsoftworks.com/products/space-wizard/how-it-works/) [Next No account. No sync. No telemetry.](https://royalsoftworks.com/products/space-wizard/how-it-works/nothing-reported/) [All of Space Wizard how it works](https://royalsoftworks.com/products/space-wizard/how-it-works/)