ace_4.9c saceserver on solaris?

Dave Matthews matthews at greengenes.cit.cornell.edu
Wed Sep 26 04:42:56 EST 2001


Hi Ed, Jing,

> > - regarding to the line in etc/inetd.conf:
> > 
> >   If I use db_nick_name at the beginning of the line,
> >   e.g.: cabbagepatch stream tcp wait ...,
> >   there is no any error msg but the client is just hanging there.
> > 
> >   If I use port_number/1 at the beginning of the line, which is like a
> > correct way for solaris tcp/rcp,
> >   e.g.: 20001/1 stream tcp wait ...,
> >   I got error msg (see below).
> 
> OK, "20001/1" is an rpc style entry, saceserver does not use rpc

Actually the db_nickname syntax doesn't leave the client hanging forever.
After some minutes it finally responds: 

  FATAL ERROR - (acedb on greengenes) reported by program saceclient (ACEDB
  4_9c), in file sclientlib.c, at line 269: Client aborted, server did not
  reply correctly during handshake

In fact ps() shows that the server never started during that time.  And
nothing gets written to database/serverlog.wrm, even if DEBUG is turned on
in serverconfig.wrm.

Note, this is only on Solaris.  In Linux this is working fine.  Identical
setup as far as I can see.

The other error message,

  FATAL ERROR - (acedb on greengenes) reported by program saceclient (ACEDB
  4_9c), in file acesocketlib.c, at line 284: connect error system error 146 -
  Connection refused

, the one you get with the "20001/1" syntax, is the same as you get if
you tell saceclient to connect to a port that isn't listed in inetd.conf or
services at all.

- Dave

---





More information about the Acedb mailing list