[TriLUG] Weird SSH behavior

Jon Carnes jonc at nc.rr.com
Mon Sep 15 16:35:40 EDT 2003


I've only seen this kind of behavior in older versions of Secure Shell,
where the daemon didn't check for dead connections. Back then I wrote a
small shell script (kicked off by a cron job) to check for the unused
connections and remove them.

Jon Carnes

On Mon, 2003-09-15 at 14:20, Tarus Balog wrote:
> Okay, on Red Hat Linux 9, two boxes, I have the following issue.
> 
> Over time, the number of sshd processes will grow. When it reaches ten  
> processes (most of them defunct), sshd will stop taking connections  
> (per the default).
> 
> Has anyone else seen this?
> 
> It may actually be an OpenNMS issue. Chris Hedemark complained that our  
> SSH poller never sent a version number, so it would cause a number of  
> logs to be generated. I modified the poller to send the version string,  
> and the logs went away, but I wonder if there is something else I need  
> to do.
> 
> I am not positive this is the case, because as I watch the process list  
> I can see the processes that OpenNMS starts, and they eventually clear.
> 
> -T
> 
> ________________________________________________________________________ 
> ___
> Tarus Balog							Main:		+1 919 545 2553
> Blast Internet Services, Inc.			Fax:			+1 919 542 5955
> Email: tarus at blast.com					URL: http://www2.blast.com/tarus
> PGP Key Fingerprint: 8945 8521 9771 FEC9 5481  512B FECA 11D2 FD82 B45C




More information about the TriLUG mailing list