CVSUP problem on SUN ULTRA 10 with CURRENT help!

[Chris Demers]

On Sun, 7 Mar 2004 14:20:12 +0000, Matthew Seaman wrote
> On Sun, Mar 07, 2004 at 07:29:09AM -0500, Chris Demers wrote:
> 
> > I have been having a problem trying to get CVSUP working correctly on a SUN
> > ULTRA 10 (256MB 40GB HDD) with CURRENT, i can ftp in the sources and do full
> > build worlds and kernel without any problems, but even after i blow away all
> > the src, ports, and sup, directories and it start recreating them when i run
> > CVSUP it gets stuck and has a weird retry time.   Any suggestions?  I have
> > tried over a dozen other CVSUP servers with the same exact results.  Even
> > using my own local one on my network does the same thing.  Only way i  have
> > been able to get sources back into the machine is to FTP them in from a snap
> > server.  Any suggestions would be apprechiated.  Note, other machines on the
> > exact same network segment have absolutely no problems using CVSUP.
> 
> Hmmm... looks like you've found a bug, probably.  All you're seeing 
> is your machine failing to connect to the cvsup server: the 'reconnect'
> message is entirely normal for that circumstance. Can you confirm 
> that you can connect to the cvsup port on any of the cvsup servers you're
> using via telnet?
> 
>     % telnet cvsup.example.com 5999
> 
> You should be able to see the banner that cvsupd produces -
>  something like:
> 
>     OK 17 0 SNAP_16_1f CVSup server ready

sun# telnet cvsup5.freebsd.org 5999
Trying 198.182.76.34...
Connected to ozone.blackened.com.
Escape character is '^]'.
OK 17 0 SNAP_16_1f CVSup server ready

Looks fine to me.  CVSUP is the only application not working on the box.  I
have reformatted it a few times so far with everything from the releases to
the JPSNAPS.

> 
> If not, then you need to work out what's wrong with the
> networking/firewalling setup on your Ultra10.
> 
> If you can connect, then there is clearly a bug in the version of
> cvsup you're using on that machine.  You should report what you're
> seeing to the port maintainer, John Polstra, (who also happens to be
> the author of cvsup) according to the instructions at
> http://www.cvsup.org/faq.html#bugreports -- in this case, as well as
> providing the data asked for there, a network dump of the traffic
> between your machine and the server would probably be useful too:
> 
>     # tcpdump -i fxp0 -w /tmp/cvsup-session.out port 5999
> 
> (replace 'fxp0' with the name of your principal ethernet interface)
> and then in another window, run your cvsup command:
> 
>     # cvsup -g -L 2 supfile
> 
> Kill the cvsup session after it says "Will retry", and then kill the
> tcpdump process.  If the cvsup-session.out file isn't too large, then
> you can uuencode it and include it in your bug report (together with
> the commands used to generate it), otherwise just mention that it is
> available if required.
> 
> Be prepared to receive patches and apply them to the cvsup sources 
> and recompile and run whatever tests John may request.

I will have to try that tonight, going out to the gym right now.  I have
looked briely even at the code and can't see anything out of the ordinary
myself, it works fine on all of my other 8 x86 boxes (from 486 to dual XEON,
running a mix of 4.X and 5.X) :)  YAY patch testing  :)

I will e-mail the port maintainer first before filing a PR, but if anyone else
has any suggestions they would be apprechiated.

> 
> 	Cheers,
> 
> 	Matthew
> 
> -- 
> Dr Matthew J Seaman MA, D.Phil.                       26 The Paddocks
>                                                       Savill Way
> PGP: http://www.infracaninophile.co.uk/pgpkey         Marlow
> Tel: +44 1628 476614                                  Bucks., SL7 
> 1TH UK

Thanks Matthew for replying, very much apprechiated.

--
Chris Demers
admin at govital.net
www.govital.net
www.govitalhosting.com