bug-cvs
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Weird bug involving :ext: access


From: Adrian Pepper
Subject: Re: Weird bug involving :ext: access
Date: Wed, 12 Feb 2003 19:49:27 -0500 (EST)

> From derek@ximbiot.com  Wed Feb 12 13:00:44 2003
> From: Derek Robert Price <derek@ximbiot.com>
> Organization: Ximbiot <http://ximbiot.com>
> To: Adrian Pepper <arpepper@mfcf.math.uwaterloo.ca>
> CC: mdb@cvshome.org, bug-cvs@gnu.org
> 
> Adrian Pepper wrote:
> 
> >Yes "cd /home/cvsuser" hangs too.
> >
> >% cd /home
> >% ls
> >
> >does not hang.  But after that, a subsequent:
> >
> >% cd xxxxx
> >
> >does hang.
> 
> I get the same hang in both cases, without an automounter involved,
> then the same host not found error.  The first time I tried, the hang
> lasted longer.  The second time CVS returned sooner, probably because
> RSH found that the DNS had cached the nonexistance of xxxxxxx.yyy.net.

By "hang" I meant forever.

The problem is actually that our automountd processes keep dying.

If automountd has died on your system, you can get the phenomenon.

The thing is that in the situation we are discussing, there seems to be
no real reason for cvs to be examining the local version of that
pathname, since it is supposed to be a remote pathname.

I.e. in the case of:

cvs -t -d :ext:cvsuser@xxxxxx.yyy.net:/home/cvsuser/cvsroot checkout cs498/a2

There seems to be no reason to expect to be able to open any of the
components of /home/cvsuser/cvsroot (presumably some sort of failure is
handled okay as part of the "checkout" operation), since they are just
specifying the remote pathname, not a local pathname.


Adrian Pepper
Math Faculty Computing Facility
University of Waterloo, Waterloo, Ontario, Canada
arpepper@math.uwaterloo.ca




reply via email to

[Prev in Thread] Current Thread [Next in Thread]