Recently one of our failover SRSS4 servers (running on OpenSuSE 10.3) ended
up having a catastrophic hardware failure necessitating rerolling a new
one.  In the past we have used a method for doing this that hasn't given us
any problems.  Basically we would do a minimal install of OpenSuSE 10.3 onto
a new machine and then pull a tar.gz backup from one of the machines on the
same cluster which included /bin /data (that's probably specific to our
heirarchy), /etc, /lib, /opt, /root, /sbin, /share, /srv, /tftpboot, /usr,
and /var.  Then it was just a matter of untarring it over the top,
reconfiguring the network services to reflect the appropriate nfs mounted
directories, hostname, and IP, and then fixing the /etc/fstab for the local
machine.  Of course then we had to reconfigure the sun ray services with
utadm and utconfig.

As I said, previously this worked without any issues.  Unfortunately this
time we've run into some problems.  The entire configuration of the sunray
services seemed to occur without any mishap, until it attempted to add 18
ldap entries to an ldap directory.

At this point the error message dealt with not being able to insert them
into the directory.  Here is where I think the problem lies.  Early on in my
time on these clusters, I was set to making the machines authenticate over
LDAP.  During this time I knew next to nothing about SRSS and had no idea
that it used LDAP.  While I was configuring the servers I noticed slapd
running on the machines (can't remember if it was some or all) and I
disabled the service.  I believe there are still backups of the original
slapd.conf file, and /var/lib/ldap still shows a large database, but I
cannot get slapd to run on any of the machines in order to determine if this
is the problem or not.

So what I'm really asking is a multipart question:  First off, what is the
role of LDAP in SRSS, and is this why the process is failing?  Second, is
there any sort of special configuration of slapd that I need to be aware of
when trying to get this service running again on the localhost?  Third, do I
need to wipe my ldap.conf files on these machines so that they can properly
speak to their own slapd?  Finally, is there anywhere that I could find an
existing configuration of this sort in order to rebuild any configuration
files that I may have destroyed?

I really appreciate any help that you can possibly offer; pointers to
resources, direct advice, anything...  The system has been working for so
long since I disabled those slapd services that I had no idea anything could
have been broken; we've added new users and had them using the SunRays,
nothing has looked broken up until this point.

There is also the possibility that some files were dropped from our tarball,
too.  So if LDAP couldn't have caused this then I know that we'll need to
drop down to single user mode in able to try a tarball that didn't have any
files locked.  According to the other admin that's been here longer, though,
this was never necessary before.

Thanks again.

Damon Getsman
_______________________________________________
SunRay-Users mailing list
[email protected]
http://www.filibeto.org/mailman/listinfo/sunray-users

Reply via email to