"Ryan P. Matijcio" wrote:
> 
> Well changing the masq timeout values doesn't to have done anything.

When having ftp problems, it's time to dust off telnet and
brush up on your ftp commands.  You can simulate the ftp client
and get some feedback and see what's working and what's not.

You may need to determine if the command connection is timing out
and if it still works when you have the problem.  It'd help if
you had some cut and paste output of the exact problem and a
repeatable test case.  I have an oxygen distro I can test from.


> Unfortunately I can't seem to find a status screen that will tell me the
> current timeout values so I can verify they have been changed. 

I couldn't find how to get it to tell you the values.
If you set it, you might try a brute force grep through
the proc directory, with care :), and seach for those
values.

I found somewhere that the  -M -S tcp tcpfin udp  style
command defaults to  15min, 2min, 5min.

Are your transferring many small files in rapid succession
or many big files?  How big?




> But as
> far as the seawall configuration goes, they are set and the firewall
> script has been restarted.  From the testing I've been doing the same
> problem is apparent.

Again please post a repeatable test case and description if at
all possible.



 
> One strange thing I've noticed is that doing a netstat -M doesn't make
> any output at all when I know it should.  Weird.

Not even after getting on a masq'd internal computer and surfing the net?
Really?


 
> Anyways, if anyone has any suggestions I'd love to hear 'em.
> 

> I've started to notice FTP problems with an Oxygen firewall I have
> running.  The problem occurs on both unix and NT systems located
> behind the firewall.  I have found that all ftp transfers be they 
> incoming or outgoing eventually freeze.  Although it appears that 
> NT -> NT transfers are more stable, they too freeze after about 
> 800-1500 files have been transferred.


After a certain number of files?  Repeatable?
Are you doing PORT or PASV transfers?

Check that you're not running out of masq ports.
I think there's 4096 ports available on a LEAF
box for masqing outgoing connections.

The ftpd, if serving incoming PASV request from
behind a masq'ing firewall, will only have a certain
number of ports dedicated to the data transfer connection.
If you use too many of them in rapid succession, then some
of them may not be completely "unESTABLISHED", iirc they
may go into some time-out wait state.  I had this exact
problem when I allowed less than 100 ports for incoming
ftpd data connections to my masqd ftpd.   I also recall
that the newer oxygen handled the reuse of ports much
better and that the problem may have gone away.  I think
it was part of the kernel upgrade in the new distro.

So maybe lowering your tcpfin or examining what's
going on in netstat -an.



> One thing I've noticed that I think is interesting to note is that this
> problem does not occur when using LeechFTP.   


Force leech ftp into debugging mode and check what's it's doing
against the ftpd server's syslog to determine what commands,
if any, it's sending on the command channel during a download.
Or do you mean it's sending NOOPs when it's idle?

Also a tcpdump of the LeechFTP commands might be interesting 
if you can't get the info from the syslog.



>I believe this because
> LeechFTP tends to keep sending commands to keep the connection active.
> That's why I was wondering if it could have something to do with the
> masq timeouts.

_______________________________________________
Leaf-user mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/leaf-user

Reply via email to