On 02/25/2011 05:36 AM, Adam Carter wrote:
> Did you remember to restart the sshd on your amd64 system? If not, try
> disabling the new high performance stuff on the client. I had a problem
> logging into a solaris box until i disabled it. In my ~/.ssh/config file I
> added;
>
> host <ip or hostname>
> HPNDisabled yes
>


I don't have a ~/.ssh/config. Will it default to ssh_config in /etc/ssh
or do I need to create one?

dhk writes:

> On 02/25/2011 05:36 AM, Adam Carter wrote:
> > Did you remember to restart the sshd on your amd64 system? If not, try
> > disabling the new high performance stuff on the client. I had a problem
> > logging into a solaris box until i disabled it. In my ~/.ssh/config
> > file I added;
> >
> > host <ip or hostname>
> >
> > HPNDisabled yes
>
> I don't have a ~/.ssh/config. Will it default to ssh_config in /etc/ssh
> or do I need to create one?

~/.ssh/config overrides settings in /etc/ssh/ssh_config.

BTW, ~/.ssh/config is very convenient so simplify access to hosts you acess
frequently.
Example: With this entry, a 'ssh -p 12345 itsme@foo.long.domain.org' can be
done simply as 'ssh foo'.

Host foo bar
HostName foo.long.domain.org
Port 12345
User itsme

Wonko

On Friday 25 February 2011 12:09:38 dhk wrote:
> On 02/24/2011 08:08 PM, dhk wrote:
> > On 02/24/2011 06:30 PM, Mick wrote:
> >> On Thursday 24 February 2011 21:51:56 dhk wrote:
> >>> Thanks, but I've tried that. ssh'ing to the hostname and loopback
> >>> address work. However, when I go out to the WAN it doesn't. So I
> >>> can't ssh user@123.123.123.123 even though I have port 22 open on the
> >>> switch for my ip.
> >>
> >> Just to state the obvious, have your tried something like:
> >>
> >> $ nc -v -z 123.123.123.123 22
> >> 123.123.123.123 (ssh) open
> >>
> >> from a WAN client to make sure that the port is open?
> >
> > I don't have the nc comand. What package is it in?
>
> The nc command does nothing when run from the same host I'm trying to
> ssh in to. I'm not sure what that means other than it can't make a
> connection. With the -w option I get:
>
> # nc -w 3 -v -z 123.123.123.123 22
> ool-43505ef2.dyn.optonline.net [123.123.123.123] 22 (ssh) : Connection
> timed out

OK, this proves that you have a connectivity problem. If the port was open
and there was a ssh server listening on port 22 for connections you would get
an "open" response.

BTW, nc is not necessary, you could use e.g. telnet to check the connection:

$ telnet 123.123.123.123 22
SSH-2.0-OpenSSH_5.8p1-hpn13v10lpk

If instead of the above you are getting no response and you are confident that
you have not defined something other than port 22 for connections to your ssh
server, then there is perhaps a firewall problem?
--
Regards,
Mick

On Tue, Feb 22, 2011 at 6:43 AM, dhk <dhkuhl@optonline.net> wrote:


After a recent upgrade to ssh I can no longer log into my Gentoo box

(amd64) from another Gentoo box (x86) that has also had a recent upgrade

to ssh. *However, I can log in to it from Suse and Redhat boxes.



Any ideas?



Thanks



dhk




While this may not help you now, one of the things Gentoo devs recommend, after updating the ssh daemon and restarting it, is to immediately try to ssh back in and make sure you still can. *Do not close your original ssh connection when you do this. *


I've had these instructions save me a few times; since I was still connected, I was able to fix the sshd_config file.

The nc command does nothing when run from the same host I'm trying to


ssh in to.*

Ok so you may not have an ssh problem (so ignore the ssh specific stuff for now) you have a network problem. It will be either routing or firewalling. If you can ping the box, then its a firewall problem. So, try pinging it first, and if that works then you know that routing is in place and its a firewall problem. If that doesn't work try traceroute to see how far you can get, and the last hop may provide clues as to why you can access it by sending an ICMP message. Post back what you find.

On Saturday 26 February 2011 07:49:44 Adam Carter wrote:
> > The nc command does nothing when run from the same host I'm trying to
> > ssh in to.
>
> Ok so you may not have an ssh problem (so ignore the ssh specific stuff for
> now) you have a network problem. It will be either routing or firewalling.
> If you can ping the box, then its a firewall problem. So, try pinging it
> first, and if that works then you know that routing is in place and its a
> firewall problem. If that doesn't work try traceroute to see how far you
> can get, and the last hop may provide clues as to why you can access it by
> sending an ICMP message. Post back what you find.

Depending on configuration of routers and firewalls ICMP packets may be
dropped, so if plain ping/traceroute fails use httping and tcptraceroute (or
traceroute -T -p 22)
--
Regards,
Mick

On 02/26/2011 03:13 AM, Mick wrote:
> On Saturday 26 February 2011 07:49:44 Adam Carter wrote:
>>> The nc command does nothing when run from the same host I'm trying to
>>> ssh in to.
>>
>> Ok so you may not have an ssh problem (so ignore the ssh specific stuff for
>> now) you have a network problem. It will be either routing or firewalling.
>> If you can ping the box, then its a firewall problem. So, try pinging it
>> first, and if that works then you know that routing is in place and its a
>> firewall problem. If that doesn't work try traceroute to see how far you
>> can get, and the last hop may provide clues as to why you can access it by
>> sending an ICMP message. Post back what you find.
>
> Depending on configuration of routers and firewalls ICMP packets may be
> dropped, so if plain ping/traceroute fails use httping and tcptraceroute (or
> traceroute -T -p 22)

I don't know why I would have a firewall or network problem, the set up
I have has been here for 8+ years. The setup is like this. In the
basement the cable internet comes in and into a cable modem. Then an
RJ45 out of the cable modem into an 8-port NETGEAR Router/Switch.
Upstairs is a hub with three computers connected and this hub is
connected to the switch in the basement. The only problem I ever had
was when the dhcp address changed, then it needed to be added to the
PORT FORWARDING section of the switch with port 22.

Remember I can still log in remotely from Redhat and Suse boxes that
weren't updated with the new openssh.

When ssh'ing in from a remote updated Gentoo box the "Read from socket
failed: Connection reset by peer" message is displayed. This message is
not displayed when trying to ssh in from inside the network.

Alright, back to the task at hand. When I tell the switch to "Respond
to Ping on Internet WAN Port" the ping statistics are good: 3 packets
transmitted, 3 received, 0% packet loss, time 2007ms.

I'm not sure where to go from here.

Is there anything in the sshd_config or ssh_config files that I need?
After the upgrade the new files were merged with the current.

Thanks
dhk

When ssh'ing in from a remote updated Gentoo box the "Read from socket

failed: Connection reset by peer" message is displayed. *This message is

not displayed when trying to ssh in from inside the network.

Ok that's different (and not consistent with the Connection timed out message, but lets ignore that for now)... try the HPN stuff i posted earlier, or even better restart sshd on the remote box.

Saturday, June 16, 2012, 3:46:14 AM, Chris wrote:

> On Fri, Jun 15, 2012 at 05:49:17PM -0700, rikona wrote:
>> ssh stops working. I initiate the ssh contact in terminal; connects
>> reliably. I can ^C to stop it OK. I sometimes get a "broken pipe"
>> message, but can reconnect easily from the CLI.
>>
>> The problem is sometimes [more often than not] ssh just quits
>> working - no message in terminal, and ^C will NOT do anything.
>> Terminal is not responding to keyboard input. If I close the
>> terminal window it warns me that something is still running and it
>> will stop if I close terminal. I usually just close the window and
>> start ssh again in a new terminal window.
>>
>> What might be happening to ssh? What is the best way to diagnose
>> this? Best of all, is there a way to keep this from happening? It
>> has become a PITA. Before the last few updates, I do not recall
>> this happening. I run 10.04, updated.
>>
> It sounds a bit like there's some sort of timeout in the connection.
> I used to suffer from this when I used ssh to connect through the
> firewall at work (with permission!). The firewall would drop the
> connection after several minutes of inactivity. Setting any/all of
> the keepalive settings in ssh didn't help, I finally resorted to a
> background script started from .profile which sent a NUL character
> every few minutes.

I'd be interested in that script...

> The symptoms of the firewall dropping the connection were exactly as
> you describe, connection stops working but no error message or other
> symptoms, just silence.

Did terminal stop responding to any input?

Thanks,

--

rikona


--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users

On Tue, Jun 19, 2012 at 2:56 PM, rikona <rikona@sonic.net> wrote:
> Saturday, June 16, 2012, 9:32:01 AM, Avi wrote:
> Sometimes it just stops working, with no error msg, as does terminal.
> If I close terminal it warns me something [presumably ssh] is still
> running and will be stopped if I exit. I exit and start over, and a
> new ssh connects to the remote comp OK. Is that also what happens with
> a disconnect, or is something else going on?

That sounds suspiciously like network issues to me... Does it happen
to anyone else on the same network?

--
Regards,
W. Scott Lockwood

--
ubuntu-users mailing list
ubuntu-users@lists.ubuntu.com
Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-users