Tectonic installer fails with ssh: handshake failed: ssh: unable to authenticate, attempted methods

575 views
Skip to first unread message

enexis

unread,
Jan 25, 2018, 5:47:11 AM1/25/18
to CoreOS User
Hi all,

Having a trouble with tectonic installer (1.8.4-tectonic.3 (0670ea7)) on bare metal after the provision:

Provision of boxes completed properly, I'm able to ssh into the controller and worker nodes with "core" user, but after starting the tectonic installer (darwin as I got a Mac) getting the following error in tectonic installer page :
Error applying plan:

1 error(s) occurred:

* null_resource.etcd_secrets: 1 error(s) occurred:

* ssh: handshake failed: ssh: unable to authenticate, attempted methods [none publickey], no supported methods remain

Searching with google reveals a few issues like, however none of is my case.
Having double checked again:
- I'm able to log in to the boxes with "core" user
- Have ssh-agent installed and key listed (just this one)
- Found a log file in /tmp/ folder which says :

2018-01-24T19:22:05.693+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:05 connecting to TCP connection for SSH
2018-01-24T19:22:05.732+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:05 handshaking with SSH
2018-01-24T19:22:05.961+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:05 handshake error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none publickey], no supported methods remain
2018-01-24T19:22:05.961+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:05 Retryable error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none publickey], no supported methods remain
2018/01/24 19:22:07 [TRACE] dag/walk: vertex "provisioner.file (close)", waiting for: "null_resource.bootstrap"
2018/01/24 19:22:07 [TRACE] dag/walk: vertex "null_resource.kubeconfig[2]", waiting for: "null_resource.etcd_secrets"
2018/01/24 19:22:07 [TRACE] dag/walk: vertex "null_resource.kubeconfig[0]", waiting for: "null_resource.etcd_secrets"
2018/01/24 19:22:07 [TRACE] dag/walk: vertex "null_resource.bootstrap", waiting for: "null_resource.kubeconfig[2]"
2018/01/24 19:22:07 [TRACE] dag/walk: vertex "provider.null (close)", waiting for: "null_resource.bootstrap"
2018/01/24 19:22:07 [TRACE] dag/walk: vertex "root", waiting for: "meta.count-boundary (count boundary fixup)"
2018/01/24 19:22:07 [TRACE] dag/walk: vertex "provisioner.remote-exec (close)", waiting for: "null_resource.bootstrap"
2018/01/24 19:22:08 [TRACE] dag/walk: vertex "null_resource.kubeconfig[1]", waiting for: "null_resource.etcd_secrets"
2018-01-24T19:22:08.961+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:08 connecting to TCP connection for SSH
2018-01-24T19:22:09.000+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:09 handshaking with SSH
2018-01-24T19:22:09.225+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:09 handshake error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none publickey], no supported methods remain
2018-01-24T19:22:09.225+0200 [DEBUG] plugin.terraform: file-provisioner (internal) 2018/01/24 19:22:09 Retryable error: ssh: handshake failed: ssh: unable to authenticate, attempted methods [none publickey], no supported methods remain

How this can be fixed or where should I look/search for ?

Thanks!
Best regards,

Kyle Brown

unread,
Jan 25, 2018, 11:59:26 AM1/25/18
to enexis, CoreOS User
Be sure that you have populated your ssh-agent before running terraform. 


--
You received this message because you are subscribed to the Google Groups "CoreOS User" group.
To unsubscribe from this group and stop receiving emails from it, send an email to coreos-user...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

enexis

unread,
Jan 25, 2018, 12:11:22 PM1/25/18
to CoreOS User
Hi Kylie,

Thanks but have double checked and it has been populated and shows up with ssh-add -l ...
Other idea how to debug it ?

Thanks,

enexis

unread,
Jan 26, 2018, 3:41:08 AM1/26/18
to CoreOS User
Hi,

For anyone who facing the same issue, main issue and fix was:
- I'm using iTerm2 for terminal, had multiple tabs open
- Have started the tectonic installer in different tab (not on that tab where I verified the ssh-add -l ...)
- Adding the ssh key with ssh-add KEYNAME in a same tab and starting tectonic installer again in that tab convinced the installer to start properly

So in Mac OS iTerm2 the ssh-add will add the key only to that tab where is launched, hence why you need to start the installer in the same tab...

Thanks Kylie to pointing out.

Best regards,
Reply all
Reply to author
Forward
0 new messages