Fernando de Oliveira wrote: > --- Em seg, 8/4/13, Bruce Dubbs escreveu: >> /usr/sbin/sshd -p29 -D -d -d > > Done.
>> On the client: >> >> ssh -v -v -p29 server > > I did some work during some time in the machines, no disconnection. > Left them alone, came back, client disconnected. From the sever log, you can see that the connection completed ate the line: debug1: Setting controlling tty using TIOCSCTTY. There were frequent delays and keep alive requests. I don't know what the ~/.ssh/config or /etc/ssh/ssh_config files had, if anything. Both those are on the client. You need to have in one of the *client* config files: ServerAliveInterval 45 The number may vary, but 45 seems to work for me. I also add ServerAliveCountMax 10 for true connection problems. Note that there is also a TCPKeepAlive setting, but the default is yes and it sometimes is not enough. There are also ClientAliveCountMax and ClientAliveInterval settings that can be set in the server, but I prefer setting the client unless there are known network problems/speed issues for the server. The questions I would have include What is the route between the server and client? Are there a lot of devices or is it as short as a virtual device in a local system? Is the route busy with other traffic? > Client lost connection with message: > > $ Write failed: Broken pipe > > Server stopped manually, after that. That's normal with a -D option at disconnect. >> ping generally uses ICMP, not tcp or udp. > I knew it could be done, was perhaps going to do it, but I want to > ping, as it is the way I "see" the other machines without going > there, or just to test they are running, etc. Generally not bad. It could be a target for a Denial of Service attack, but otherwise it won't do much. If someone sends an ICMP packet (ECHO REQUEST), it just responds with an ECHO_REPLY packet. -- Bruce -- http://linuxfromscratch.org/mailman/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
