Skip to content

FTP Server (hosted by heavendata)

This article describes how to connect to our ftp server to access import and export files.

The server accepts plain FTP and encrypted FTP (FTPS). Use FTPS: over plain FTP, your password and your files cross the internet unencrypted. Connect to ftp.eu.40three.net, the name the server’s certificate is issued for, so that a client checking the certificate accepts it. The encryption is explicit, on port 21, not implicit: nothing answers on port 990, so do not use ftps:// addresses or implicit encryption.

WinSCP is a Windows application that allows you to manually connect to our ftp server and browse available files.

You can download it here: https://winscp.net/

In the login dialog, choose FTP as the file protocol and TLS/SSL Explicit encryption as the encryption.

Curl allows you to download files from ftp servers on command line. This is perfect if you want to write a script to automatically download and import data.

Install Curl:

Now you can download files using a single command

curl --ssl-reqd ftp://ftp.eu.40three.net/CHANNEL_ID/products.csv --user USERNAME:PASSWORD -o products.csv
# example
bin/curl.exe --ssl-reqd ftp://ftp.eu.40three.net/a28ca68c-bbb8-4753-ab35-30e83980ab7e/products.csv --user myuser:secretpassword -o products.csv

--ssl-reqd makes curl encrypt the login and the file transfer, and stop with an error rather than fall back to plain FTP. Do not use --ssl instead: it falls back to plain FTP without telling you.

The username and password are the ones set in the channel’s Publishing step. The folder is named after the channel’s id. It is the only folder there after you log in, so a desktop client shows you its name.

Import folders and how long files are kept

Section titled “Import folders and how long files are kept”

Each data source that receives files over FTP has a folder of its own at the root of the server, and two folders inside it. Browse the root in an FTP client to see the exact names — a data source appears under its key, or under its id if it has none.

Path What is in it
<data source>/import Where you drop files.
<data source>/archive/<timestamp>-<id> One folder per import, holding the files that import read.

When an import starts, the files it takes are moved out of import into a new folder under archive. That is why a file you uploaded is no longer in import — it has been read, not lost.

A file stays in import until every file the data source expects has arrived, so a partial set waits there for the rest.

Enable a new data source before you upload to it. A data source is created with Enabled unchecked, and its FTP login works only once you have enabled it. Until then the server refuses the login.

Archived imports are kept for 30 days, or the most recent 5,000 imports of that data source — whichever comes first. After that the folder and its files are removed. The import folder is yours: nothing clears it automatically, so a file that never completed a set stays until you remove it.

Runs themselves, with their logs, are listed in the app for the same 30 days — see Background jobs.

  • Can I use the default Windows command line ftp client? No. This client does not support passive FTP and will not work in most environments. Active FTP means, that our ftp server would have to initiate data connections and this would be blocked by your firewall / router.