Do your transfer rate measurements include login time? If so, have you tried an interactive ftp to see how long the login is taking? If it is taking a long time, the login time by itself may explain your slow transfer rate, while the transfer itself may be working just as fast as a transfer to Windows. We had slow Linux ftp logins (30 seconds before we even got a prompt for the userid) which turned out to be due to three reasons: the ftp server on Linux was running under xinetd, not as a daemon; xinetd was issuing ident requests (port 113) back to the client machine, due to USERID being specified in /etc/xinetd.conf or the /etc/xinetd.d file for the ftp server; and port 113 traffic was blocked. I won't go into more detail, since I don't know if any of this applies to your situation.
Bill On Mon, 16 Jun 2008 15:14:52 -0400, François Paré wrote: >Hello, > > > >I'm running a batch job that does a FTP transfer from a z/OS mainframe to a LINUX server and the transfer rate is about 20K/sec. If I do the same FTP transfer to a Windows server I got a transfer rate of about 900K/sec. The mainframe OSA card is running at 100 Mb/sec full. Since the Windows and the LINUX server are on the same switch and got the same throughput capability I suppose that there is an optimal setting that is done automatically when the transfer is done with a Windows server but this setting is not done automatically with a LINUX server. I tried PASV but it didn't change the bad transfer rate. Could you tell me what this setting could be? Thank you! > > > > > >********************************************************* *** > >Francois Pare > ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
