[Dxspider-support] Cluster Hang

[Mike Lewis]

Thanks to everyone who sent me a reply. This has been a NAGGING issue for a while. I FINALLY solved it on the eve of the CQDX WW SSB contest, and was able to stay connected to my cluster throughout the contest. I am posting my findings just in case someone else has an issue like this. It was NOT a Spider issue.

The problem was some software that was installed on both of the PC laptops I traditionally use to remote connect to my Linux system. I sometimes log into our work network from home using my laptops. Therefore I have some Cisco VPN software installed. It turns out this software was the problem. It would drop the connection to the spider computer after a certain amount of inactivity. This would happen even if I was not using the VPN software to connect to my workplace. Turns out the Cisco VPN software runs a service at boot time on the Windows PC that is the culprit. Once I disabled that, the link was completely bulletproof. Thinking back after the fact, the time period during which I first had this installed coincides with the start of my troubles.

So if anyone is experiencing problems keeping a remote session alive to a spider system from another Windows box, it just might be your VPN software!



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

> -----Original Message-----
> From: dxspider-support-bounces at dxcluster.org 
> [mailto:dxspider-support-bounces at dxcluster.org] On Behalf Of 
> Bela Markus
> Sent: Thursday, October 25, 2007 12:33 AM
> To: The DXSpider Support list
> Subject: Re: [Dxspider-support] Cluster Hang
> 
> Hi Mike,
> 
> the leading number is the standard UNIX time interpreted as 
> seconds elapsed from January 1st, 1970.
> 
> I don't think your issue is related to SUSE. The very high 
> CPU load usually caused by corrupted dupe and/or user file, 
> same happened to me also after a migration to new hardware. 
> Delete dupe first and restart spider. If you are changing 
> distribution, for server I strongly advice CentOS.
> 
> Yes, a chat.
> 
> Regards... Béla
> 
> 
> Mike Lewis írta:
> >
> > I am still having all kinds of problems getting my clkuster to run 
> > reliably, mostly since switching to running on a SuSE 10.2 
> distro. I 
> > have had problems with my local telnet sessions hanging, 
> but now have 
> > an even bigger problem where the cluster itself seems to hang. I 
> > recently (within the last week) loaded a new build of DXSpider (it 
> > shows as V1.54 build 0.172).
> >
> > Here is an excerpt from the log showing the last entries 
> prior to the 
> > hang. I had logged on, composed a message to a friend, 
> logged out and 
> > then later back on, and then left myself connected. I had been 
> > expanding the usdbraw file to add state info, but had not 
> yet run the 
> > load/usdb command. When I came back to the system, the 
> terminal that I 
> > had run the client program in was not returning any prompts. top 
> > showed the cluster.pl using a consistent 95% or more of the cpu.
> > killing the client and re-running did not connect to the cluster. I 
> > had to manually kill the cluster.pl instance.
> >
> > Log file:
> >
> > 1193281900^ann^ALL^IZ7AUH-6^IZ7AUH-6 DX CLUSTER -> 
> dx.iz7auh.net port 
> > 8000 1193282956^msg^msg 1 from KE0MF to KB0TVH stored 
> > 1193283206^DXCommand^KE0MF disconnected 1193283284^DXCommand^KE0MF 
> > connected from 127.0.0.1 1193284847^ann^ALL^PA4JJ-2^dx 
> cluster telnet 
> > pa4jj-no-ip.org port 8000
> > 1193285786^chat^MW^SM3BEI^#49 vaken!
> >
> > A few questions:
> >
> > What is the way to interpret the leading number (I am 
> assuming it is a 
> > time stamp of some kind?) on each line of the log. Is it 
> possible to 
> > determine from the log the amount of time between entries?
> >
> > What is the last entry telling me? is this some sort of a 
> chat request 
> > to my node?
> >
> >
> > If there is anyone out there with DXSpider experience on SuSE 
> > (preferably with a relatively new version) maybe they can 
> help me. I 
> > had this running on a Debian distro on an older box with less 
> > problems. I guess I could scrap SuSE and try going back to a Debian 
> > (or some other Linux) install, but I have other unrelated 
> reasons for 
> > wanting to keep this system running SuSE.
> >
> > ML
> >
> >
> >
> >
> > --
> > This message has been scanned for viruses and dangerous content by 
> > *MailScanner* <http://www.mailscanner.info/>, and is believed to be 
> > clean.
> > 
> ----------------------------------------------------------------------
> > --
> >
> > _______________________________________________
> > Dxspider-support mailing list
> > Dxspider-support at dxcluster.org
> > http://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
> >   
> 
> 
> _______________________________________________
> Dxspider-support mailing list
> Dxspider-support at dxcluster.org
> http://mailman.tobit.co.uk/mailman/listinfo/dxspider-support
>