Petre Scheie wrote:
I've noticed for several months that when I ssh into a k12ltsp box from a remote location, that while I get connected quickly, it takes about 60 seconds before I get a prompt. I've traced the problem to the call to ltspinfo in /etc/profile.d/ltsp.sh which is sourced during login. Here's the line: eval `/usr/bin/ltspinfo -h $LTSP_CLIENT -c all 2> /dev/null` Since I'm coming from a remote location, rather than a client, I'm guessing that $LTSP_CLIENT, which holds the address of where I'm coming from, and therefore is not part of the client subnet, is causing the trouble; the problem does not occur when I ssh in from a machine on the same local network as the server. Is this command trying to talk to ltspinfod on my remote machine? Since I'm not a thin client, it doesn't exist, and presumably the above eventually times out waiting for an answer that will never come. But that doesn't explain why I don't have the same problem when I connect from a non-thin-clint machine that is on the same subnet. Anyone seeing this problem? Petre
You are correct that ltspinfo tries to make a connection to ltspinfod on the remote machine. I'm guessing that your firewall drops packets rather than rejects them. Ltspinfo will keep on trying to connect for up to 60 seconds. If it does get a reject, it stops trying immediately. You can change this by editing /usr/bin/ltspinfo and changing: my $timeout = 60; to something shorter. -Eric
Attachment:
signature.asc
Description: OpenPGP digital signature