[nos-bbs] JNOS 2.0d - 1 year anniversary, INP3, and other stuff

[Barry Siegfried]

[maiko at pcs.mb.ca wrote]:

> Hello Ruben,
>
> > After i do a domain query Jnos hangs and exits to shell.
>
> I just tried it as well, and noted the following :
>
> 1) domain query <ip address>
>
> works okay.
>
> 2) domain query <domain name>
>
> seems to hang, then eventually I get disconnected from my
> telnet session to JNOS (I used sysop mode from BBS), BUT
> the JNOS is still running and working.
>
> > What could be happening?
>
> I don't know yet. It could be something as simple as the 'domain'
> parameters.
>
> Barry, got any ideas ?

A few, although they ask more questions than they answer.  Unfortunately,
a detailed search through my release notes since the beginning didn't
find a reference to this specific problem with 'domain query'.

I have done lots of work on the domain client resolution code over the
years so my model is not so similar to the JNOS model anymore.  I've
added some functionality which JNOS may or may not have, and I've
bulletproofed some items in the domain client code, particularly as
they relate to the session and process pointers for the current query
process.  Of note, I have put a backwards link to the session struct
in the proc struct of the current 'domain query'.  This means that
if the 'domain query' is aborted or alarm()ed then it will be able
to quickly find the associated session to free.  This may or may
not have something to do with the problem that is occuring (although
I somehow doubt it).

It sounds as if the crash is happening either when the DNS returns the
reply, or when the timeout wait for a reply is over.  This leads me
to the question:  Has anyone trace()d the channel and determined
whether this is happening when a DNS reply is received or rather when
the timeout wait for a reply is over?

I have a few other questions:

1) The form of the command in JNOS *is* 'domain query <dns hostid>
   <hostid> [<type> [<maxtries> [<timeout (ms)>]]]' isn't it?
2) If #1 is yes, then what happens when you try other types of
   record queries, like cname, mx, soa, txt, etc.?
3) If #1 is yes, then *what machine* is the <dns hostid> from
   which it is trying to obtain the resolution?
4) Is this happening in a program that has the domain nameserver
   code compiled in and running?  This theoretically shouldn't matter
   unless *this machine* is the <dns hostid> for *this machine*
   (an unlikely configuration I would imagine).
5) Is this happening in a program that does NOT have the domain
   nameserver code compiled in?

The more things we can observe and answer, the better the chance
we have of narrowing down the source of the problem.

All of this reminds me to ask this (off this topic):

Is anyone using 'domain translate on'?

73, de Barry, K2MF >>
           o
          <|>      Barry Siegfried
+---------/-\---------------------------+
| Internet | bgs at mfnos.net              |
| HomePage | http://www.mfnos.net/~bgs  |
+----------+----------------------------+
| Amprnet  | k2mf at nnj.k2mf.ampr.org     |
| PBBS     | k2mf at k2ge.#cnj.nj.usa.noam |
+----------+----------------------------+