NetNews Usenet Archive 1992 #31/NN_1992_31.iso/spool/comp/unix/aix/12766

Text File | 1992-12-22 | 2.7 KB | 52 lines

Newsgroups: comp.unix.aix
Path: sparky!uunet!zaphod.mps.ohio-state.edu!swrinde!gatech!darwin.sura.net!mojo.eng.umd.edu!russotto
From: [e-mail] (Matthew T. Russotto)
Subject: Re: Weird Ping Problem -- Hide 'n' Seek?
Message-ID: <[e-mail]>
Date: Tue, 22 Dec 92 03:06:01 GMT
Organization: Project GLUE, University of Maryland, College Park
References: <[e-mail]>
Keywords: tcp ping
Lines: 40

In article <[e-mail]> [e-mail] (Ed Ravin) writes:
>I've got the strangest problem trying to communicate between two RS/6000's.
>They're networked to each other via token ring and a couple of bridges.
>
>Let's call these two machines A and B.  Everything was wonderful until one
>day, some connections and configurations were moved around.  I signed on to
>machine A and tried to ping machine B.  No response.  I called up the admin
>of machine B on the phone.  He said, "Let me try pinging you..."  He signs on
>to machine B and pings me.  It works fine.  I then go back to machine A
>and try pinging B again.  Much to my amazement, now it works, but after several
>minutes (haven't nailed down the exact length of time it takes) if I don't
>reference machine B in any way, it fails again.
>
>The admin from B was skeptical, so I came down to his machine and started up
>two X windows.  The first one I left at the regular command prompt, and the
>second one I telnetted to machine A, but via two indirect hops so as not to
>create a direct path between B and A.  From that window, I said "ping B", and
>it hung there, timing out.  I then switched to the local window on B, with
>the sysop watching, and said "ping A".  Immediately, both windows began
>scrolling with successful pings.
>
>Anyone out there have some ideas?  The folks in charge of the token ring
>bridges claim everything works, and that traffic should pass equally in
>both directions.  I've looked through B's routing tables, but I can't find
>anything that explains why locally originated pings would work while remote
>ones go unanswered...

B has a static route to A.  A has no static route to B, but is set up
for dynamic routes.  But there's no router around telling A where B
is, so A doesn't know B.  But when B pings A, A figures out a route to
B by looking at the incoming packet (clever, ain't it?), so things
work, until that route expires.  If the routes are supposed to be
static, you should be able to add a static route to B on A and have
everything work.

--
Matthew T. Russotto    [e-mail]    [e-mail]
Some news readers expect "Disclaimer:" here.
Just say NO to police searches and seizures.  Make them use force.
(not responsible for bodily harm resulting from following above advice)