Re: URGENT**** (fwd)

Charles Scott (cscott@freeway.net)
Thu, 2 Nov 1995 11:31:55 -0500 (EST)

Phil:
OK, let me be a little more clear about this. Our symptoms, which I
believe are consistant with the other reports I've had, were that there
were no other network problems than the one host could not send a packet
to the one PortMaster. There were no configuration changes before the
incident and none after aside from the update to 3.1.4. I'm not saying
that I'm convinced it was the PortMaster which caused this, however, it's
very suspicious. If it had only been the one wierd occurance on our
system, I would have discounted it as an effect of cosmic rays or
something, but when someone else came up with the same wierd symptom it
really perked my ears up.
Yes, I made a fatal syntactical error and refered to an ethernet
interface as a "port" which may have been cofused.

Chuck Scott, Pres.
freeway, Inc.
cscott@freeway.net

On Thu, 2 Nov 1995, Phil Jensen - News Administrator wrote:

>
>
> On Wed, 1 Nov 1995, Brian 'MegaZone' Bikowicz wrote:
>
> > I completely disbelieve we caused it though any fault of ours.
>
> On Thu, 2 Nov 1995, Charles Scott wrote:
>
> > Brian:
> > I think you should be carefull about making such statements. We had
> > this happen to us also, just after updating a PM to 3.1.4. There is
> > nothing I can find in our host or PM configurations which could have
> > caused the symptoms we saw. We have never seen this happen with anything
> > else, and it only affected the hosts ability to talk to the PM, nothing
> > else! (RIP updatess?????)
>
> Just because there is nothing -you- can find, doesn't mean it is a RADIUS
> bug, or a problem with your Portmaster. Did you check your interface(s)?
> What operating system are you using? You can find out quite a bit of
> information about your shell host by just typing:
>
> # ifconfig <interface>
>
> For example:
>
> sierra# ifconfig ef0
> ef0: flags=a863<UP,BROADCAST,NOTRAILERS,RUNNING,SIMPLEX,LINK1,MULTICAST>
> inet 205.199.144.101 netmask 0xffffff00 broadcast 205.199.144.255
>
> It's not a port on the UNIX host, that gets trashed, either. He said he
> was having problems with his network all-around, and not just a single
> port. I've had the same problems myself about a year and a half ago, but
> they were of my doing, and not because of a Portmaster.
>
> > Like the others, our PM sent the authentication request to our RADIUS
> > server, but the RADIUS server was unable to get the reply back to the
> > PM. Our logs show that the authentication request was received and the
> > response sent many times for each connection attemp, but the PM never
> > received it. Since you don't need to do anything to the PM to clear the
> > problem, it's obvious that it's the port on the RADIUS host that gets
> > trashed. The question is why!
> > It sure will be interesting to see a routes table from a host that's
> > having this problem.
>
> Agreed.
>
> > Chuck Scott, Pres.
> > Freeway, Inc.
> > cscott@freeway.net
>
> Phil Jensen _\\|//_
> News Administrator (-0-0-)
> ---------------------------------------------------------------ooO-(_)-Ooo----
> ValleyNet Communications - Central California's Premier Internet Provider
> ------------------------------------------------------------------------------
> Voice: (209) 486-VNET (8638) 2300 Tulare, Suite 100
> Fax: (209) 495-4940 Fresno, CA 93721-2226
> Data: (209) 495-4950 http://www.valleynet.com
> ==============================================================================
>
>