Matt Johnston wrote:
On Tue, Dec 08, 2009 at 03:47:02PM -0800, Ahilan Anantha wrote:
Hi List,
I plan to use "dbclient" as a low memory footprint alternative to
OpenSSH's "ssh" for SSH tunnels.
On the client I have software that creates SSH tunnels to many systems.
Sometimes the connection to these remote systems will break, at which
point "ssh" will exit. The exit gets detected and the connection gets
reestablished. But this works in "ssh" because I'm using the
ServerAliveInterval and ServerAliveCountMax options. Without them, ssh
would never check that the connection was up and I'd have to wait an
eternity for a TCP timeout. Or implement my own heartbeat on top of the
tunnel.
dbclient sends an "ignore" packet every N seconds, but I
don't think that elicits a server response. It will
generally time out after a minute or so when the client OS
gives up on receiving an ACK, though SIGSTOP is a funny
case since the remote OS is probably still sending TCP ACKs.
I'll take a look at implementing something closer to what
ServerAliveInterval does (sending something that will fail
and checking for a reply, iirc).
OpenSSH's "tcpkeepalive" just sets the TCP keepalive option
on the socket with setsockopt(), but won't probe the
connection itself.
Cheers,
Matt
Thanks, Matt.
OpenSSH's client is sending an "SSH2_MSG_GLOBAL_REQUEST" with a bogus
request type of "[email protected]" with want reply set to 1. And on
the server side it doesn't try to match that name and just always sends
an "SSH2_MSG_REQUEST_FAILURE" when it gets that message.
And then every time the client gets an SSH2_MSG_REQUEST_SUCCESS or
SSH2_MSG_REQUEST_FAILURE it sets the number of outstanding server alives
to 0.
Regards,
Ahilan
