I am hoping anyone of you could help me in this weird problem we have in 1 of our Solaris 10 servers. Lately, we have been having some ftp problems in this server. Though it can ping any server within the network, it seems that it can only ftp to a select few. For most servers, the ftp connection either times out or is rejected. All IP configurations in the server are correct, as verified by our network team. All system configurations on ftp seem to be working as shown below:
Yes we did. ftp is working fine on the destination servers we tested (outbound and inbound).
What's more strange in this server having problem is that all inbound ftp connections are working fine. It is only its outbound connection that results to rejection/timing out.
A few things to check from the basic to the more obscure.
1) Can you ftp within the server both to localhost and the server name?
ftp localhost
ftp localservername
2) Is the server name and IP address consistent?
Check on every machine:
nslookup servername1
nslookup servername2
3) Are there rogue .netrc files ?
Search each source computer for .netrc files and inspect permissions and contents.
4) FTP can be crippled by certain physical network configuration issues
No network port on the server, switch or hub should be set to auto-detect or auto-negotiate. The mess this can create is very difficult to detect but a common symptom is connections dropping to half-duplex and CRC errors.
A cold start of network components and then servers will often cure these symptoms for a while. A random order cold start (e.g. after a power failure) will often cause these symptoms. Re-patching a network connection is another way of causing these symptoms.
5) FTP does not like the "jumbo packets" setting in any network components and will often hang.