Network Timeout issue

thead · Post by **thead** » Wed Dec 10, 2008 5:56 pm

Not convinced this is a CoreFTP issue yet, but I thought I'd see if anyone else has had this;
I have the server running at a remote site and I set it up (second time - first being last week in testing) and configured to my liking, mostly default settings but some have been adjusted to *relax* restrictions.

I have two users created and I can connect and move files, all looks to be good. Sometimes though for no apparent reason I will not be able to connect and get a "Network error: Connection timed-out" message.
I'm using both WinSCP and CoreFTP client with exactly the same result.

I can also telnet to port 22 and get the CoreFTP version number when all is well. When it isn't it comes back unable to connect.

Not sure what's causing it as nothing will change between when it works and when it doesn't. Though it seems to work/fail in waves where it will work from three different client desktops just fine and then stop working for no apparent reason.

Restarting the service *does not* resolve, nor does running it as an application.

There is a firewall between them but port forwarded and since I can connect sometimes that seems to be an unlikely culprit.
Could be a networky issue but I have seen no consistency between when it works and when it doesn't.

Any help?

thead · Post by **thead** » Wed Dec 10, 2008 7:19 pm

I'm currently using the defaults again (10 and 3) but had bumped them up to 99 and 99 just to test.

I've also upped the # of connection entries in the per user settings to no avail.

I have since installed the server on a second box within the same subnet just to test and I'm getting the same results, even with a fresh config and when connecting from within the subnet.
So there shouldn't be any fwall issues...

Very strange.

thead · Post by **thead** » Wed Dec 10, 2008 7:24 pm

Another thing - even on the new server I cannot reproduce the error locally (that is to connect from an FTP client on the server to the server itself always works).

I had this on the first server I tried as well.

thead · Post by **thead** » Wed Dec 10, 2008 7:50 pm

I suppose I should mention that I'm running Server 2K3 R2 SP2 64 bit.
Can't find anything on the site about which Windows versions are supported...

Testing this issue on some other machines...

thead · Post by **thead** » Thu Dec 11, 2008 2:47 pm

Thanks. I'm still fighting with it but so far I'm not getting any love from the DEP changes.
I'm afraid if I can't get this working soon I may need to look at other apps out there. I was hoping to buy Core as it seems quite solid - save this issue.

Thanks for your help so far, and any additional tips out there would be greatly appreciated.

thead · Post by **thead** » Thu Dec 11, 2008 5:07 pm

Not an issue here. My first instance had two total users but I don't think I ever had both connected simultaneously.

The second instance has only one user and I've only been testing Core for about a week now.

thead · Post by **thead** » Thu Dec 11, 2008 6:01 pm

I may have found it, and it may not be a Core issue at all.
The environment we're trying to deploy it in is in a Windows NLB. Our plan had been to have the server coexist with the NLB but have it's own dedicated IP address, but it looks like when you manually add an IP address to a node in an NLB that node gets the IP address added to it's NLB Clustered IP address list but no other nodes get it.

Still testing this out but once I removed one of the servers from the NLB and started Core on that server everything was fine and dandy (so far).
Another oddity was that while one machine had the FTP IP the other two nodes in the NLB could not ping that IP. Once removed from the NLB the Core IP address was pingable by all three.

Not sure how I'm going to resolve this for our deployment of Core but it's looking more and more like NLB is to blame.

I don't suppose anyone has ever successfully set up Core in an NLB environment?