Problem fxping

[Kid Twilight]

I have a weee problem fxping. Whenever I transfer a file from a IoFTP server to a RaidenFTP it just stops when the file is done and just sits there wating for the next file...

226 Transfer finished successfully. Data connection closed .

Then I have to hit abort and start again in order to fxp the next file...

Cheers for any help

*EDIT* Im not running the server, i just promised to look into the problem...

[darkone]

ioFTPD performs graceful socket shutdown which might cause up to 6minute delay, if other side is not posting shutdown notification.

Most common implementation is:

...
shutdown(SD_BOTH);
closesocket();

Some servers are doing (wrong implementation, which may cause ioftpd to wait until timeout):

...
closesocket();


ioFTPD does (optimal implementation :)):

setsockopt(); // No lingering
...
shutdown(SD_SEND);
recv();
closesocket();

[Kid Twilight]

Thanks for the reply darkone. So I should take this to the raiden forum and ask?

[acid6trik]

ioFTPD 5.1.3r
windows all versions
problem when fxping from io to raiden


Hi all,

I have try to contact raiden developpers, and they answer me that they don't care at all about ioFTPD server and that raiden works fine with all other ones ... They don't seem to like your server ^^
Furthermore I don't understand why can we fxp from the 4.9.x to raiden, but not from the 5.x ? You should have change something in the code between this twos, as said juste before, and can you reverse this change for the next versions please ? ;) or produce two versions in that way ? ;)
Because raiden is used yet somewhere, and I would like to use my v5 but while this problem is, I can't.

Otherwise, what a great work with ioFTPD you did ! :banana:
thanks darkone for this excellent server ;)

[Superblue]

rather than have this graceful shutdown that doesn't interoperate for shit, can you take it back out or make it configurable?

[darkone]

It allows server to free resources quicker + uses less resources (as kernel doesn't have to manage lingering) I'm not quite sure, if 5.1.3 has the bug-fixed implementation.. should have tho :)