>>>>> "ED" == Edward Doolittle <[EMAIL PROTECTED]> writes:

    ED> On 27 Feb 2000, Jake Colman wrote:

    >> Well, this may be the final problem.  A central part of fixing the
    >> "lost packet" problem is to flush the named cache by doing a 'ndc
    >> restart' whenever diald drops the link.  When I try this, however, the
    >> 'restart' always retriggers diald.

    >> Feb 27 11:31:25 firewall diald[19903]: Trigger: udp 207.198.222.7/53
    >> 192.168.0.100/1430

    ED> Hmmm, looks like your named is sending from port 1430, not port 53.
    ED> Make sure the query-source address line is functioning correctly in
    ED> your named.conf , and make sure named is reading the correct conf
    ED> file.

It seems that it cycles through random 143x ports.  I'm really not clear on
what this is all about.  How do I check whether the 'query-source address'
line is functioning correctly and how do I confirm that named is using the
named.conf file that I think it's using?

    ED> If everything is fine with query-source address, then your named must
    ED> be sending some other chatter out on high ports.  Perhaps there's
    ED> another option to stop that.  I dont have that problem.  Maybe try
    ED> commenting out the forwarders line in named.conf , and any other line
    ED> mentioning 207.198.222.7 .

But why would it be chattering at all?

In any event, I think I know what to do about all this.  If I understand
correctly, the goal here is to 1) ensure that named does not bring up the
link spuriously and 2) ensure that diald is triggered by a udp packet (for
named lookup) and not a tcp packet.

Even without the 'query-source address' line in my named.conf, my named does
not seem to be overly verbose and my link does not go up unexpectedly.  I
guess that the entires in my standard.def filter are sufficient.  To ensure
that diald is triggered with a udp packet, I wrote an 'ip-goingdown' script
that does an 'ndc restart'.  Even though I mentioned in a prior post that an
'ndc restart' seems to trigger a connection, if I do it before the link goes
down it does not matter.  I realize true that an 'ip-goingdown' script is
only executed when diald terminates the link itself and not when the remote
side forces it to terminate.  That should not be a problem, however, since,
if the remote side terminates, diald automatically attempts to reconnect
anyway.

So, does any of this make sense to you or have I missed the entire point of
the exercise?

-- 
Jake Colman                     

Principia Partners LLC                  Phone: (201) 946-0300
Harborside Financial Center               Fax: (201) 946-0320
902 Plaza II                           Beeper: (800) 928-4640
Jersey City, NJ 07311                  E-mail: [EMAIL PROTECTED]
                                       E-mail: [EMAIL PROTECTED]
                                          web: http://www.ppllc.com

microsoft: "where do you want to go today?"
linux:     "where do you want to go tomorrow?"
BSD:       "are you guys coming, or what?"

-
To unsubscribe from this list: send the line "unsubscribe linux-diald" in
the body of a message to [EMAIL PROTECTED]

Reply via email to