Tracking the SLURPD trouble

[Stephen Anderson <stephena@HiWAAY.net>]

I continue to track the problem I am having with multiple replicas and
slurpd in the hopes that someone will be able to help my remedy the
situation.

After examining debugging output of slurpd (-d 68) I believe I have an
idea at what the problem may be.  The slurpd communications are stepping
on each other.

In the original problem statement, a modification proceeded to be sent to
replica-A correctly, but replica-B reported "ERROR: Value already exists".
Somehow the command sent to replica-A got intercepted and performed by
replica-B as well.  To further test this idea I changed the binddn on both
replicas to original names, Rep1 & Rep2 respectively.

Now a different set of symptoms show up.  The slupd process tries to bind
to replica-A using Repl1 immediately followed by a call to bind to
replica-B using Repl2.  Replica-A then responds that this is an unkown
object.  I believe somehow Replica-A got sent the bind request intended
for Replica-B.

This is all VERY strange.  Please, can somebody help me straighten out
this mess?  Thank you.


                                       Stephen Anderson
                                     <stephena@HiWAAY.net>