[dev] Anyone else seeing weirdness with anoncvs tonight?

[Jon Parise]

On Thu, Jun 06, 2002 at 09:52:08PM -0500, Brent J. Nordquist wrote:

> $ for x in README [ikmnt]*/README; do cvs status $x | grep Sticky.Tag; done
>    Sticky Tag:          RELENG_2 (branch: 2.4.2)
>    Sticky Tag:          RELENG_3 (branch: 2.9.2)
>    Sticky Tag:          RELENG_1 (branch: 1.2.2)
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> 
> $ for x in README [ikmnt]*/README; do cvs status $x | grep Sticky.Tag; done
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> cvs [status aborted]: recv() from server anoncvs.horde.org: EOF
> 
> ...then if I wait a few minutes I can do a few more "cvs status" before I
> start to get the recv() EOF again.  cvs.horde.org isn't doing this (though
> that's across SSH rather than pserver).

I'd wager it has something to do with the fact that the repository on
anoncvs.horde.org is NFS mounted from cvs.horde.org.  I don't have a
fast solution to the problem at the moment, but I'll keep it in my for
the next time I shuffle services around.

I may just remove the NFS mount in favor of hourly rsync's between the
two machines.  The only downside there is that the anoncvs repository
is, at most, an hour behind the live repository.

-- 
Jon Parise (jon@csh.rit.edu)  .  Information Technology (2001)
http://www.csh.rit.edu/~jon/  :  Computer Science House Member