I have couple of debian boxes at the office where we have a dns server, so ftp'ing to these boxes from a windows machine using a .net application is fairly faster and we had no issues. As closer we get to deployment of our project, we get to set up a test environment where we connect these windows boxes and debian boxes via a switch and there's no dns server, all IP addresses are manually set and they are static. And we're experiencing a very peculiar problem, all ftp operations are timing out. Machines can ping each other, we have connectivity, but it's just that ftp is super slow. Just to test the situation I went into debian command line tried to ftp the machine that I am on using it's IP address(let's say machine's ip address is 16.1.X.X)
ftp 16.1.X.X
Machine instantly responds saying connected to 16.1.x.x but it takes up to 40 seconds prompting for ftp username and password.
Since I am on the machine I also tried:
ftp 127.0.0.1
it instantly asked for username and password. So it looks like it takes too long to identify 16.1.X.X since there's no dns server.
I tried running ftp -dv 16.1.X.X
Only extra information I am receiving is ftp: setsockopt: bad file descriptor.
I didn't have this problem in my office, so I am suspecting that it has something to do with DNS, any ideas will be greatly appreciated.
I not certain if this will help, but if you have no DNS server you should populate the
/etc/hosts
file on each host.Here is an example format:
Does Debian use vsftpd? If so you could disable reverse lookups with adding this to your .conf file.