[Dxspider-support] FW: Telnet Sessions locking up

[Mike Lewis]

 Thanks Dirk, I'll give those a try next chance I get (I'm at work
now...)

ML


----------------------------------------------------
This mailbox protected from junk email by MailFrontier Desktop
from MailFrontier, Inc. http://info.mailfrontier.com

-----Original Message-----
From: Dirk Koopman [mailto:djk at tobit.co.uk] 
Sent: Monday, January 15, 2007 3:53 PM
To: Mike Lewis
Subject: Re: FW: [Dxspider-support] Telnet Sessions locking up

Mike Lewis wrote:
> Dirk,
>  
> I greatly apologize for sending this directly to you, but I just can't

> seem to get anyone to send me any meaningful feedback on this problem 
> (last posted to the reflector on 12/21/06 with no replies...). I did 
> get one response from Charlie Carroll the last time I asked him about 
> it, but he was just asking me some further questions about my 
> configuration, and did  not really have any insights to give me. In 
> any case, this does not seem like it should be a big issue to fix 
> (maybe just some config somewhere that is wrong or something) but I 
> have not been able to turn anything up. The net result is, that while 
> my users that access the node over the 2m link are quite happy, the 
> connects I make to the node locally over my internal network all seem 
> to timeout. It makes the node very frustrating to use.
>  
> Any thoughts?
>  
> Thanks bunches,
>  
> ML
>  
>  
> 
> ----------------------------------------------------------------------
> -- This mailbox protected from junk email by MailFrontier Desktop from

> MailFrontier, Inc. http://info.mailfrontier.com 
> <http://info.mailfrontier.com/>
> 
>  
> 
> ----------------------------------------------------------------------
> --
> *From:* dxspider-support-bounces at dxcluster.org
> [mailto:dxspider-support-bounces at dxcluster.org] *On Behalf Of *Mike 
> Lewis
> *Sent:* Thursday, December 21, 2006 10:17 PM
> *To:* dxspider-support at dxcluster.org
> *Subject:* [Dxspider-support] Telnet Sessions locking up
> 
> Hi,
>  
> I am running version 1.52 build 61.523, and am having an exasperating 
> problem. My node connects to one other node (usually K1EA), and my 
> Spider software which is hosted in a Linux box (running Debian) seems 
> to work fine. External connects to the cluster are restricted to be 
> via 2 meters only, and those all seem to work with no problems. I can 
> also log into the cluster when logged in  on the actual LInux box that

> is hosting my Spider node. The problem comes in when I connect to the 
> cluster over my local network using either of my 2 windows computers. 
> I usually use Putty to run a telnet session, or sometimes connect via 
> my DXLabs logging software (using the Spotcollector application). I 
> have also just run a command window under Windoze XP and connected 
> using a telnet command. In each of these cases, I connect ok, and the 
> link works fine, but if I have even a short period (say more than 
> maybe 5 mins or
> something) where I am not querying the cluster or at least hitting a 
> return key once in a while the telnet session seems to die. The 
> cluster does not seem to show me as disconnected, but I get no 
> response back from Spider with any of my telnet typing, and the telnet

> application will eventually report the connection as terminated. I can

> relog in any time after the session has frozen (after closing the 
> original session
> window) and the sysytem connects me again to the cluster as if nothing

> had happened. The log shows no activity that explains what is wrong, 
> it just seems to have nothing in it after my last interaction that was

> recognized by the cluster before the freeze. Regular telnet sessions 
> to the Linux system, without logging into Spider, don't exhibit this
behavior.
>  
> I have been running this same configuration for many years, and never 
> used to have this issue at all. It seemed to crop up about a year ago,

> when I upgraded the hardware I was using for the cluster and 
> reinstalled. I recently had a hard drive crash, and had to re-install 
> again. I was hoping this problem might go away, but alas, it is still 
> there. Anyone have any thoughts as to what the issue is? Any help 
> would be appreciated.

As it is a linux box, I would suggest that you look at things like:

  iptables -nL

to see if there is any gratuitous firewalling and then (as I suspect
there isn't any or it isn't the problem), you should run ethereal (or
wireshark as it seems now to be called), monitoring port 7300 and seeing
what happens.

You could even send me a pcap file if you like.

Dirk