cvs pserver config worked on Mdk 8.0, fails on RH 7.1...why?

[Gerald Combs]

1) Did you HUP xinetd, or run '/etc/init.d/xinetd reload'?

2) What does 'grep xinetd /var/log/messages' show?

3) What happens when you telnet to port 2401 on the server?

4a) Does the user listed in /etc/xinetd.d/cvs exist?

4b) Can you do cvs-y things as that user locally?

4c) If the user in question doesn't own the repository outright, do they
    have group perms?  Does /etc/xinet.d/cvs have a "groups = yes" line?
    
5) Have you run truss^H^H^H^H^Hstrace, e.g.

   Run 'strace -fF -o /tmp/cvs.strace -p <xinetd's PID>'
   
   Try pserver access.  Wait for it to die.
   
   Kill the strace session.  Inspect the output file, looking for stuff
   like EACCESS.

On Tue, 31 Jul 2001, Kendric Beachey wrote:

> We have just adopted CVS as our source control platform in my
> department.  When I was doing the proof of concept to convince
> everyone it was a good idea, I was (and still am) running CVS on my
> desktop machine, which runs Mandrake 8.0.
> 
> Now that CVS has been blessed, the boss has bought a fancy new machine
> to serve as our real CVS repository, and it runs Red Hat 7.1.  I have
> verified that cvs the client works just fine on the new machine, and I
> successfully used 'cvs init' to start a new repository.  The problems
> are with trying to make the pserver run on the new machine.
> 
> Like my desktop machine, the new one uses xinetd instead of regular
> inetd, which is what is used in the examples in Karl Fogel's excellent
> documentation.  The man pages gave me enough to set it up to run
> properly on Mandrake, but doing the same thing on Red Hat isn't
> working.  Here is an example of what happens when I try to run 'cvs
> login' against the new machine:
> 
> 
> [beachey at zimmerdale beachey]$ cvs -d > 
>      :pserver:beachey at cvs.garmin.com:/cvs login
> 
> (Logging in to beachey at cvs.garmin.com)
> CVS password:
> cvs [login aborted]: recv() from server cvs.garmin.com: Connection
> reset by peer
> [beachey at zimmerdale beachey]$
> 
> 
> It doesn't matter if I type the right password or not.
> 
> The relevant details of the config files:
> * /etc/services already contained tcp and udp lines for port 2401 for
> cvspserver
> * /etc/xinetd.d now contains a file called "cvs" that is identical to
> my working one on the Mandrake machine--it points to
> /usr/sbin/cvspserver
> * /usr/sbin/cvspserver is a shell script that reads /etc/cvs/cvs.conf
> to find the value(s) of CVS_REPOS, then calls 'cvs pserver' using a
> separate '--allow-root' for each item in CVS_REPOS
> * /etc/cvs/cvs.conf just says CVS_REPOS="/cvs"
> * ummm...anything else you'd like to know about, I can provide.  :-)
> 
> Many thanks in advance for any help you can give!
> 
> PS: I just noticed that sometimes when trying to do 'cvs login' from
> the actual cvs machine, I sometimes get the message "cvs [login
> aborted]: received broken pipe signal" instead of the connection reset
> message.
> 
> -- 
> Kendric Beachey
> ak at kc.rr.com
> 
> DVD decryption in seven lines of Perl code:
> $_='while(read+STDIN,$_,2048){$a=29;$b=73;$c=142;$t=255;@t=map{$_%16or$t^=$c^=( 
> $m=(11,10,116,100,11,122,20,100)[$_/16%8])&110;$t^=(72, at z=(64,72,$a^=12*($_%16 
> -2?0:$m&17)),$b^=$_%64?12:0, at z)[$_%8]}(16..271);if((@a=unx"C*",$_)[20]&48){$h 
> =5;$_=unxb24,join"", at b=map{xB8,unxb8,chr($_^$a[--$h+84])}@ARGV;s/...$/1$&/;$ 
> d=unxV,xb25,$_;$e=256|(ord$b[4])<<9|ord$b[3];$d=$d8^($f=$t&($d12^$d4^ 
> $d^$d/8))<<17,$e=$e8^($t&($g=($q=$e14&7^$e)^$q*8^$q<<6))<<9,$_=$t[$_]^ 
> (($h=8)+=$f+(~$g&$t))for at a[128..$#a]}print+x"C*", at a}';s/x/pack+/g;eval
> 
> 
>