On Mon, 2006-11-06 at 10:18 -0800, Mike Wright wrote: > John Austin wrote: > > On Mon, 2006-11-06 at 09:35 -0800, Mike Wright wrote: > > > >>Hi all, > >> > >>This is my third day of chasing it and am still trying to resolve this > >>issue. > >> > >>Left the box running overnight but by morning it had hung. Reboot found > >>a trashed format: disk completely inaccessable, so... > >> > >>Just performed a clean install of fc6 and have not added any new > >>packages or any other software. The machine was given a host name that > >>resolves with dns. Just to be sure iptables/ip6tables was disabled on > >>machines at both ends of the connection. > >> > >>I opened an xterm just beneath the clock, ssh'd to a remote machine on > >>the same subnet, then ssh'd back to the new fc6 box. It gave me back a > >>password prompt. > >> > >>At 9:00:00 I pressed enter in response to the password request > >>At 9:00:40 the message "Last login: Sun Nov 5...." > >>At 9:01:00 I finally got the remote system prompt [mike@test ~]$ > >> > >>I repeated this numerous times and each time it took exactly 60 seconds: > >>40 seconds until the "Last login" message, then 20 more until the sytem > >>prompt. > >> > >>I think this 20/40/60 seconds may be the clue, but don't know what > >>services/processes have those numbers hard coded into them. > >> > >>ssh/rsync outbound works correctly; only inbound connections have this > >>problem. > >> > >>Has anybody else experienced this? Would definitely appreciate help > >>with this. > >> > > > Just a suggestion as you are desperate ! > > Put the names and IP's in /etc/hosts on both machines and try again > > I have a feeling the delay will go > > > > DNS reverse lookup may then be the problem > > > > Thanks both Mike McGrath and John Austin for your replies. > > I have seven machines on two subnets all using dns, with nothing in > /etc/hosts except the localhost entry. Everything has been working fine > for years, including internal and external dns, smtp, and ssh. > > From the new fc6 box I can ping every other host using only its > hostname (not fqdn) so that shows that it can talk to dns. Nonetheless, > by adding the remote's name to the fc6 box's host file ssh/rsync works > correctly. Does this mean that all my fc4 and rh8 boxes are not working > correctly? > I would suspect something in DNS resolution on the machine that is delayed in responding. Maybe the resolv.conf file has an invalid nameserver entry so it has to wait for the timeout before it continues. > Perhaps the real question is, "What has changed that makes the two > versions behave so differently. fc4 has ssh-4.0 and fc6 has ssh-4.3. > Can that alone explain the difference in behaviors? > > Now I'm even more baffled than before. I guess it's time to revisit my > dns database. > > Thanks again for your help. > :m) > -- fedora-list mailing list fedora-list@xxxxxxxxxx To unsubscribe: https://www.redhat.com/mailman/listinfo/fedora-list
