[Fedora-directory-users] error while using migrate-ds-admin.pl from 1.04 to 1.1.0

[Tony]

On Sat, Apr 19, 2008 at 3:53 PM, Rich Megginson <rmeggins at redhat.com> wrote:

> Tony wrote:
>
> > I'm trying to upgrade from 1.0.4 to 1.1.0 on a CentOS 5 system.  I
> > tested everything in a vm and all went fine, but come to the real live
> > server and when I try to migrate the data, this happens:
> >
> > [root at sputnik ~]# migrate-ds-admin.pl -f temp.inf
> > Beginning migration of Directory and Administration servers from
> > /opt/fedora-ds . . .
> > Beginning migration of directory server instances in /opt/fedora-ds . .
> > .
> > Your new DS instance 'slapd-directory' was successfully created.
> > [18/Apr/2008:22:34:50 +0100] createprlistensockets - PR_Bind() on All
> > Interfaces port 389 failed: Netscape Portable Runtime error -5982 (Local
> > Network address is in use.)
> > [18/Apr/2008:22:34:50 +0100] createprlistensockets - PR_Bind() on All
> > Interfaces port 389 failed: Netscape Portable Runtime error -5982 (Local
> > Network address is in use.)
> > Could not start the directory server using command
> > '/usr/lib/dirsrv/slapd-directory/start-slapd'.  The last line from the error
> > log was '[18/Apr/2008:22:34:50 +0100] createprlistensockets - PR_Bind() on
> > All Interfaces port 389 failed: Netscape Portable Runtime error -5982 (Local
> > Network address is in use.)
> > '.  Error: Unknown error 256
> > Exiting . . .
> > Log file is '/tmp/migrateCGDfkB.log'
> >
> > Anyone seen this before, or can help me get around it?  I think the
> > migrate script is trying to start the new ldap server having not managed to
> > stop the old one- does that make sense? But the old version has to be
> > running in order to do the migration.... catch 22?
> >
> The old version should not be running.  Is there some documentation that
> leads you to believe that the old version has to be running?  If so, we need
> to fix it.


Ah, that would be it then- bit obvious really.  I don't think I saw anything
to say either way whether the services should be running or not.  I wasn't
taking enough notice on my test run and simply assumed that because the new
services were running after the migrate script had run that the old ones
must have been running beforehand - a sort of symmetry. But with standard
init scripts being a new thing in this release I guess it would be tricky to
have detected and/or stopped the old services in a reliable way within the
script.

I ran the migration again with the services initially stopped and it worked
just fine of course.  Thanks for your help.

Cheers,

Tony
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/fedora-directory-users/attachments/20080420/b172350e/attachment.htm>