How to use generated SSH-KEY in pipelines?

jai_c7sdgk

I generated a key pair in repo -> settings -> ssh keys

and copied the pub key to autorized_keys of server

and added the fingerprint on bitbucket too

 

but my pipelines fail with

 

<span>  <br>File "/usr/local/lib/python3.6/site-packages/fabric/connection.py", line 29, in opens</span><span>    <a href="http://self.open/" target="_blank" rel="noopener nofollow noreferrer">self.open</a>()</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/fabric/connection.py", line 615, in open</span><span>    self.client.connect(**kwargs)</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/paramiko/client.py", line 437, in connect</span><span>    passphrase,</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/paramiko/client.py", line 749, in _auth</span><span>    raise saved_exception</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/paramiko/client.py", line 662, in _auth</span><span>    key_filename, pkey_class, passphrase</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/paramiko/client.py", line 571, in _key_from_filepath</span><span>    key = klass.from_private_key_file(key_path, password)</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/paramiko/pkey.py", line 206, in from_private_key_file</span><span>    key = cls(filename=filename, password=password)</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/paramiko/ed25519key.py", line 76, in __init__</span><span>    data = self._read_private_key("OPENSSH", f)</span><span>  <br>File "/usr/local/lib/python3.6/site-packages/paramiko/pkey.py", line 289, in _read_private_key</span><span>    raise SSHException("not a valid " + tag + " private key file")<br></span><span>paramiko.ssh_exception.SSHException: not a valid OPENSSH private key file</span><span>File "<decorator-gen-3>", line 2, in run</span>

 

 

Do I really need to store the private key in pipelines env var ?

Can't the pipeline pick up the default identity by itself?

 

Any help is welcome

 

Graham Gatus

If you inspect the ~/.ssh/config directory in your pipeline, it contains an entry for IdentityFile, which points to the location of the private key which was generated via the UI. This location resolves to  '/opt/atlassian/pipelines/agent/ssh/id_rsa'. Using tools such as git or ssh, they would check ~/.ssh/config to find the key location.

I notice you are using a python based tool - theres a possibility it is looking for the private key under ~/.ssh/, which doesn't exist. If this is the case, you can either configure the tool to use the correct location (/opt/atlassian/pipelines/agent/ssh/id_rsa), or copy the key from /opt/atlassian/pipelines/agent/ssh/id_rsa to under ~/.ssh/ to satisfy the tools requirements.

Storing keys in environment variables would only be necessary if you required multiple keys as Pipelines currently only supports adding a single key via the UI (e.g ssh into multiple services, or using git+ssh for committing back to Bitbucket and ssh for connecting to a remote server).