info-cvs
[Top][All Lists]
Advanced

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

RE: help on cvspserver (in linux mandrake environnement) - correction -


From: Art
Subject: RE: help on cvspserver (in linux mandrake environnement) - correction - Ok now
Date: Thu, 17 Jan 2002 20:11:45 -0600

Ok, My problem was that the server had "/cvs/cvsroot" which
I renamed to "/cvs/CVSROOT" so it could reference the "passwd"
file underneath - I also did a chown -R and chmod -R to cvs:cvs
with 775 and made anoncvs a member of the group cvs.

So, it all works now. I verified this on the Linux and Windows
sides.

BTW, another note on the net said there's no need to kill xinetd
since it's a super server, presumably, it rereads the configuration
for each service request.

Art

> -----Original Message-----
> From: Art [mailto:address@hidden
> Sent: Thursday, January 17, 2002 6:46 PM
> To: address@hidden
> Subject: RE: help on cvspserver (in linux mandrake environnement) -
> correction
>
>
> Oops! Sorry - I just found out I was not executing:
>
>  cvs -d :pserver:address@hidden:/cvs update -d -P
>
> as I said below. I was using a script which does:
>
>  export CVSROOT=:pserver:address@hidden:/cvs
>  cvs update -d -P
>
> and I just realized that the CVS content overrides the
> environment variable with:
>
>  CVS/Root:
>
>   :pserver:address@hidden:/cvs
>
> When I did do the '-d' command I got,
> from Linux:
>
>  /usr/local/src/ccvs
> bash-2.05# cvs -d :pserver:address@hidden:/cvs update -d -P
> 02/1/address@hidden:08:41: DEBUG: {main_loop} select returned 1
> 02/1/address@hidden:08:41: DEBUG: {server_start} Starting service cvspserver
> 02/1/address@hidden:08:41: DEBUG: {exec_server} duping 12
> 02/1/address@hidden:08:41: DEBUG: {main_loop} active_services = 7
> 02/1/address@hidden:08:41: DEBUG: {check_flags} flag entry 0 processed
> 02/1/address@hidden:08:41: DEBUG: {child_exit} waitpid returned = 3382
> 02/1/address@hidden:08:41: DEBUG: {server_end} cvspserver server 3382 exited
> 02/1/address@hidden:08:41: DEBUG: {child_exit} waitpid returned = -1
> 02/1/address@hidden:08:41: DEBUG: {main_loop} active_services = 7
> cvs [update aborted]: received broken pipe signal
>
> From Windows Cygwin:
> $ cvs -d :pserver:address@hidden:/cvs update -d -P
> cvs [update aborted]: reading from server: Connection reset by peer
>
> which elicits this from the server (debug mode):
> bash-2.05# 02/1/address@hidden:40:37: DEBUG: {main_loop} select returned 1
> 02/1/address@hidden:40:37: DEBUG: {server_start} Starting service cvspserver
> 02/1/address@hidden:40:37: DEBUG: {exec_server} duping 12
> 02/1/address@hidden:40:37: DEBUG: {check_flags} flag entry 0 processed
> 02/1/address@hidden:40:37: DEBUG: {child_exit} waitpid returned = 3383
> 02/1/address@hidden:40:37: DEBUG: {server_end} cvspserver server 3383 exited
> 02/1/address@hidden:40:37: DEBUG: {child_exit} waitpid returned = -1
> 02/1/address@hidden:40:37: DEBUG: {main_loop} active_services = 7
>
> Sorry for the misinformation.
> So, I'm now troubleshooting my server setup!
> BTW, I previously had my repository /cvs -> <a_Windows_partition>, but
> I couldn't get a lock, so I couldn't get a lock, i.e., I was getting this:
>
> $ cvs -d :pserver:address@hidden:/cvs update -d -P
> cvs [update aborted]: reading from server: Connection reset by peer
>
> address@hidden ~/ws/ccvs
> $ cvs -d :pserver:address@hidden:/cvs update -d -P
> ? configure_ac_252
> ? configure_tmp
> ? gmon.out
> ? lib/gmon.out
> cvs server: Updating .
> cvs server: failed to create lock directory for `/cvs/ccvs'
> (/cvs/ccvs/#cvs.lock
> ): Permission denied
> cvs server: failed to obtain dir lock in repository `/cvs/ccvs'
> cvs [server aborted]: read lock failed - giving up
>
> and the server said:
>
> bash-2.05# 02/1/address@hidden:44:31: DEBUG: {main_loop} select returned 1
> 02/1/address@hidden:44:31: DEBUG: {server_start} Starting service cvspserver
> 02/1/address@hidden:44:31: DEBUG: {main_loop} active_services = 7
> 02/1/address@hidden:44:31: DEBUG: {exec_server} duping 12
>
> Linux would not let me create the anoncvs account without a password,
> however, cvsadmin didn't seem to care. Well. It doesn't work, so maybe
> it does care...
>
>
> Art
>
> > -----Original Message-----
> > From: Art [mailto:address@hidden
> > Sent: Thursday, January 17, 2002 3:03 PM
> > To: address@hidden
> > Subject: RE: help on cvspserver (in linux mandrake environnement)
> >
> >
> > > -----Original Message-----
> > > From: address@hidden
> [mailto:address@hidden Behalf Of
> > > Matt Riechers
> > > Sent: Thursday, January 17, 2002 7:47 AM
> > > To: address@hidden
> > > Cc: address@hidden
> > ...
> > > address@hidden wrote:
> > ...
> > > > I don't manage to start the cvs pserver service on the linux
> > > server, though i
> > > > think the files /etc/inetd.conf and /etc/services are well
> > > configurate  (whith
> > > > cvspserver on port 2401). I've buid a repository in a
> > /usr/local/cvsroot
> > > > directory.
> > >
> > > The relavent contents of /etc/services and /etc/inetd.conf would
> > > help diagnose
> > > the problem.
> >
> > I am also experiencing the same problem.
> > My system is 2.4.8-26mdk.
> > My services files includes:
> >   cvspserver      2401/tcp   cvspserver
> >   cvspserver      2401/udp   cvspserver
> >
> > My system does not use inetd.conf, rather it uses:
> >   xinetd.conf
> >   xinetd.d/
> > and the dir has a file called "cvs" which I thought included the
> > right stuff for cvspserver:
> > service cvspserver
> > {
> >         disable             = yes
> >         socket_type         = stream
> >         protocol            = tcp
> >         wait                = no
> >         user                = root
> >         server              = /usr/sbin/cvspserver
> > }
> >
> > However, I get this error in response to my request:
> > cvs [update aborted]: connect to a260(192.168.0.2):2401 failed:
> > Connection refus
> > ed
> >
> > The command is:
> >
> >   cvs -d :pserver:address@hidden:/cvs update -d -P
> >
> > I created anoncvs via cvsadmin and the Unix adduser commands.
> >
> > > > When i try to start the pserver manually (in /usr/sbin) i get
> > > the message "cvs
> > > > [pserver aborting] bad port auth start:"
> > If I try this, I get:
> >
> >   cvs [pserver aborted]: bad auth protocol start:
> >
> > > cvs pserver is meant to be run from inetd. What messages do you
> > > get when run
> > > from inetd?
> > >
> > > -Matt
> >
> > So, I edited the "xinetd.d/cvs"  file to say:
> >
> >         disable             = no
> >
> > vs "yes", killed xinetd and restarted it.
> > Now, my remote pserver command works!
> >
> > I wish I knew this was the right way to manage xinetd, or
> > that xinetd had a "restart" client command to direct it
> > to re-read its config files.
> > Well, at least I didn't have to reboot Linux!
> >
> > Art





reply via email to

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