The two-client test is a manual test you can perform to check your setup and familiarize yourself with the Codex API. These steps will guide you through running and connecting two nodes, in order to upload a file to one and then download that file from the other. This test also includes running a local blockchain node in order to have the Marketplace functionality available. However, running a local blockchain node is not strictly necessary, and you can skip steps marked as optional if you choose not start a local blockchain node.
Optionally, if you want to use the Marketplace blockchain functionality, you need to also include these flags: `--persistence --eth-account=<account>`, where `account` can be one following:
Next replace `<SPR HERE>` in the following command with the SPR returned from the previous command. (Note that it should include the `spr:` at the beginning.)
Notice we're using a new data-dir, and we've increased each port number by one. This is needed so that the new node won't try to open ports already in use by the first node.
We're now also including the `bootstrap-node` argument. This allows us to link the new node to another one, bootstrapping our own little peer-to-peer network. (SPR strings always start with "spr:".)
Normally the two nodes will automatically connect. If they do not automatically connect or you want to manually connect nodes you can use the peerId to connect nodes.
curl -X GET http://127.0.0.1:8080/api/codex/v1/connect/<PEERIDHERE>?addrs=/ip4/127.0.0.1/tcp/8071
```
Alternatively on Mac, Linux, or MSYS2 you can run it in one command like:
```bash
curl -X GET http://127.0.0.1:8080/api/codex/v1/connect/$(curl -X GET -H "Accept: text/plain" http://127.0.0.1:8081/api/codex/v1/peerid)\?addrs=/ip4/127.0.0.1/tcp/8071
Notice that we are sending the peerId and the multiaddress of node 2 to the `/connect` endpoint of node 1. This provides node 1 all the information it needs to communicate with node 2. The response to this request should be `Successfully connected to peer`.
### 5. Upload The File
We're now ready to upload a file to the network. In this example we'll use node 1 for uploading and node 2 for downloading. But the reverse also works.
(Hint: if curl is reluctant to show you the response, add `-o <FILENAME>` to write the result to a file.)
Depending on the file size this may take a moment. Codex is processing the file by cutting it into blocks and generating erasure-recovery data. When the process is finished, the request will return the content-identifier (CID) of the uploaded file. It should look something like `zdj7WVxH8HHHenKtid8Vkgv5Z5eSUbCxxr8xguTUBMCBD8F2S`.
### 6. Download The File
Replace `<CID>` with the identifier returned in the previous step. Replace `<OUTPUT FILE>` with the filename where you want to store the downloaded file.
Notice we are connecting to the second node in order to download the file. The CID we provide contains the information needed to locate the file within the network.
### 7. Verify The Results
If your file is downloaded and identical to the file you uploaded, then this manual test has passed. Rejoice! If on the other hand that didn't happen or you were unable to complete any of these steps, please leave us a message detailing your troubles.
