| 1 |
On 06/12/17 06:43, Alan McKinnon wrote: |
| 2 |
> On 06/12/2017 00:35, Ian Zimmerman wrote: |
| 3 |
>> On 2017-12-06 05:53, Bill Kenworthy wrote: |
| 4 |
>> |
| 5 |
>>> No, all machines are set up as keyless ssh - git has never needed it |
| 6 |
>>> there. In frustration I created keys and set portage up as a keyless |
| 7 |
>>> ssh account as well, no change. |
| 8 |
>> |
| 9 |
>> ssh messages are sometimes misleading. For instance, ssh would say |
| 10 |
>> something like "pubkey authentication failed" when in fact I prohibited |
| 11 |
>> root logins on the server. |
| 12 |
>> |
| 13 |
>> I'd try connecting with bare ssh as the user in question, with maximum |
| 14 |
>> verbosity turned on (-vvv). |
| 15 |
>> |
| 16 |
> |
| 17 |
> |
| 18 |
> The error messages from the ssh client are, by design, intentionally |
| 19 |
> vague. They amount to a teeny bit more detail than just "something went |
| 20 |
> wrong", plus the available auth methods listed in parenthesis. |
| 21 |
> |
| 22 |
> This is because the sshd server avoids information leakage that |
| 23 |
> attackers could use. |
| 24 |
> |
| 25 |
> To find out why ssh does not work, start by looking at the server logs, |
| 26 |
> then examine the client is nothing obvious stands out. |
| 27 |
> |
| 28 |
|
| 29 |
Got it! Needed ssh keys for portage@remote from root@local. Its working |
| 30 |
but no idea why its only this machine that required it. |
| 31 |
|
| 32 |
Thanks, |
| 33 |
BillK |
Archives