[Info-vax] stumped by SSH

[David Froble]

Jan-Erik Soderholm wrote:
> Den 2016-02-06 kl. 17:34, skrev Phillip Helbig (undress to reply):
>> In article <n94rpv$o9p$1 at dont-email.me>, Simon Clubley
>> <clubley at remove_me.eisner.decus.org-Earth.UFP> writes:
>>
>>>> ssh -V says:
>>>>
>>>> local:  3.2.0 on Alpha
>>>> remote (linux): OpenSSH_6.7p1
>>>
>>> So now we know the remote system is a non-VMS box, you need to up the
>>> debugging level on the VMS SSH client and see if the debugging messages
>>> give you any clues.
>>
>> How can I do that?
>>
>>> You also have not said yet whether both the working and non-working
>>> accounts are on the same Linux box.
>>
>> Working and non-working refer to VMS accounts.  Some VMS accounts work
>> and some don't, no matter from which account (local or remote) one tries
>> to log in.
>>
>>> There was an issue where increased
>>> SSH security defaults in recent Linux versions means that the VMS TCP/IP
>>> services client is not considered secure enough and hence connections
>>> from the VMS client are rejected.
>>
>> I know about this problem, but it cannot explain why I can log in
>> (locally or remotely) to some VMS accounts but not to others.
>>
>>> I am now seriously confused. Is the remote SSH server a VMS box or
>>> a Linux box ?
>>
>> Local: VMS cluster.  Remote: Linux.
> 
> I think that Simon implied that:
> 
> "Local" is where you are trying to start a SSH session *from*.
> That is, the envoronment where you are when issuing the SSH command.
> Or in other words, the "SSH client".
> 
> "Remote" is where you are trying to connect *to*.
> Where the "SSH server" is running.
> 
> If we do not agree on that, this will be realy weird... :-)
> 
> Jan-Erik.
> 
> 
> 
> 
>   Some VMS accounts cannot be logged
>> in to via SSH, neither locally nor remotely.  Telnet works locally and
>> remotely and SET HOST/LAT works locally.  So, it is not a problem with
>> the account (wrong password, etc).  Nor is it a problem with SSH per se,
>> as some VMS accounts work fine.
>>
> 

Philip,

If things are happening differently in two different user accounts, then they 
are not the same.  Something is different.  Possibly some case sensitive stuff?

If you are having a problem with a captive user account, make it non-captive 
during testing.  You might get a bit more information.

I'm not really familiar with the break-in evasion, however, it's possible that 
it could be based on user account, in addition to other things.  Actually, I'm 
rather sure it is.

YOu don't post much detail, and as Steven may have mentioned in the past, not 
too many of us are mind readers or physic ....