What happened is that FTP and SSH services do a reverse DNS query to ensure that hosts actually are who they say they are, when your connection went out, they were no longer able to do this, but keep trying anyway, slowing things down to the point of frustration or timeout.
The fix is simple, but makes your connections a little less secure, as this check won't be performed on all connections, not a big deal if your boxes only face internal networks.
First, (assuming you're using openssh) fix SSH:
echo "UseDNS no" >> /etc/ssh/sshd_config
Next (assuming your're using pure-ftpd) fix FTP:
echo 'yes' > /etc/pure-ftpd/conf/DontResolve
Restart both services through /etc/init.d or upstart and you're set!
No comments:
Post a Comment