Shell Programming and Scripting

I have a file with size as 120,371 bytes in local machine and trying to FTP it to a remote server using put command. say that it transferred exactly the same size of 120,371 bytes and hence it returns a 226 code.

I have a doubt on how the rule is made for 226 - if both size on local and remote are equal only it returns 226 ?. And what will be the code return if it transferred is lesser than 120,371 bytes might be due to any interruption ?

226 is not a program exit code, it is a FTP status code and means closing data connection - requested file action successful.

See: List of FTP server return codes - Wikipedia, the free encyclopedia

Yes, I do meant the return code after a FTP command.

so on what basis it returns 226 which in turn means "closing data connection - requested file action successful".

You're missing the point. When you run an ftp command, it connects to an ftp server on the target system. When you terminate the ftp session, the (local)ftp command will tell the (remote) ftp server to close the connection.

I haven't looked at the internals of the handshakes between ftp and the ftp server. The ftp server might return a status code to ftp for each requested operation or it might return a consolidated status code to ftp when the connection is closed. I would expect the former rather than the latter (and I would expect the ftp server status code to be used by the ftp utility use that status code to indicate whether or not each request (e.g., get, mget, or put) succeeded or failed. Whether or not ftp sets its exit code based on the server status report(s) isn't specified on the ftp man page on OS/X (and presumably BSD) systems. I would expect a non-zero exit status from ftp if the ftp server reports failure on any (or maybe just on the last) requested operation.

The ftp utility is not contained in the POSIX standards so I can't give you a specification of what the standards require in this area.

The status messages returned during an ftp session (sometimes requires the -v flag) do not affect the final return code of the ftp client command. They are useful to check for things such as failure to change remote directory, permissions to write a file denied etc.

Unless ftp fails to connect to the host at all (before a login attempt even) you will always get a zero return code to the ftp command.

The messages you are seeing are status messages. Those from 000 to 399 are considered as normal good responses and those of 400-999 are failures, although not all the possible messages have been used. If you capture the standard output (that would normally come to the screen) in a log file, then you can scan it for these messages. I have two files that I use for this. One looking for records the start with three digits over 400, and another to exclude text such as 512 bytes sent successfully and I do all my error checking in the shell script that calls the ftp.


I hope that this clarifies things a bit.



Robin

Although ftp response codes are not part of POSIX, they are standardized by an RFC:

https://tools.ietf.org/html/rfc640

This explains the meaning of each of the digits in the 3-digit response code.

As rbatte1 noted not all of them are used on every implementation. And some do not make sense for certain implementations.

FWIW: ftp has one major security flaw: usernames and passwords are sent over the connection as plaintext. So ftp is probably okay for use inside a firewall.

Actually, the range is 400 to 599, not 400-999, where:

Code:

     4xx   Transient Negative Completion error-codes
     5xx   Permanent Negative Completion error-codes

The ABNF for an error-response is:

Code:

      error-response = error-code SP *TCHAR CRLF
      error-code     = ("4" / "5") 2DIGIT