Dartmouth · Geisel School of Medicine ← All groups

Downloading your data

Pick a transfer method based on where you're sending the data, not who you are. The destination drives the decision: rsync only helps if both endpoints are on Dartmouth research-computing storage; for a laptop, lab server, or any off-site machine, use wget over HTTPS regardless of whether you're an Internal lab member or an External collaborator.

Click-to-download in the browser

The fastest path for one or two files: open the per-project page on the portal, then click any item in the Contents list. Folders open into the auto-index (still password-gated); files trigger save-as via the browser. This works for everyone — Internal users via Dartmouth CAS, External users via the per-lab Basic auth password emailed to you.

For larger pulls (more than ~10 GB total, or many files at once), use one of the terminal methods below. They're resumable, verify integrity, and don't depend on the browser staying open.

rsync — preferred when copying within DartFS / Discovery / Andes

If you're SSH'd into a Dartmouth research-computing host (Discovery, Andes, a Polaris VM) and want to copy data into your home directory or a lab share on DartFS, rsync over SSH is the right tool. It's faster, resumable, and verifies integrity. Requires a Dartmouth NetID — External collaborators without SSH access should skip to the wget section.

Replace <NETID> with your NetID, <Lab> with the lab folder name (the slug shown in the portal URL), and <project> with the project subdirectory:

rsync -avhP --info=progress2 \
  <NETID>@andes8.dartmouth.edu:/dartfs-hpc/rc/lab/G/GSR_Active/<Lab>/<project>/ \
  ./<project>/

You'll be prompted for your NetID password. Run from the destination machine (typically a login node, lab server, or your Discovery home dir).

wget — command-line download to anywhere else

Use this when the destination is your laptop, a lab server outside DartFS, an off-site cluster, or any other system that isn't on Dartmouth research-computing storage. Works for both Internal users (NetID via Dartmouth CAS) and External users (per-lab username + password).

External users (per-lab Basic auth)

Your username is the lab name (the slug in the URL — e.g. Adimab, Murawala). The password is the random string in the email from fred.w.kolling.iv@dartmouth.edu. The password is shared across your lab; treat it as a lab credential.
wget --user=<Lab> --ask-password \
  --recursive --no-parent --no-host-directories --cut-dirs=2 \
  --reject="index.html*" \
  https://rcweb.dartmouth.edu/GSR_Active/<Lab>/<project>/

It'll prompt for the password. Resumable across runs (just re-run the same command). Output lands in ./<Lab>/<project>/.

Internal users (NetID via Dartmouth CAS)

The portal uses Dartmouth CAS for Internal labs, which doesn't play well with wget directly. Two options:

curl (single file, External users)

curl -u <Lab> -O https://rcweb.dartmouth.edu/GSR_Active/<Lab>/<project>/<file>

Verifying the transfer

Each project that's been processed by the GSR will eventually have an md5sums.txt file at its root. The "Download md5sums" button on the per-PI page links to it (or shows "md5sums pending" if it hasn't been computed yet — the nightly job catches up). After download:

cd <project>
md5sum -c md5sums.txt

Trouble?

Reply to the credential email or contact fred.w.kolling.iv@dartmouth.edu.