So I've noticed that when I reboot my Windows server running the active
directory remote loader, it doesn't want to re-connect back to the
engine, running on a NetWare box. I see the following traces below:

On the Engine (on NetWare 6.5SP7):
-Message: Code(-9006) The driver returned a "retry" status indicating
that the operation should be retried later. Detail from driver: No
connection to remote loader-

On the Remote loader side:
-Loader: Waiting for DirXML to connect on 'TCP server socket, port
XXXX, address AAA.BBB.CCC.DDD'...-


That's pretty much the relevant trace data -- The remote loader is on
trace level 4, and the engine is on trace level 5, and the engine only
has that one error in yellow coloring (the other messages are just
rehashes of the driver making another attempt at a connect and aren't
relevant).

I found last time that I ran into this that I had to change the port
from 8090 to 8091 on both sides, as well as reboot my NetWare server.
Attempting to unload DIRXML.NLM hung java up something bad (killed all
services running on the server, but left the console responsive), and I
was forced to cold boot the machine to get it running again, which I
really hate doing to a NetWare box, despite NSS being pretty robust.
That got it all running again.

But I decided to reboot my Windows server last night, being that it IS
a Windows server and thus not designed for staying online for long
periods of time (and there were a bucket load of security patches to
apply anyways -- yay!) like a NetWare or Linux/Unix server is. And
well, after the reboot, the remote loader and the engine aren't seeing
each other again. I half expect the two to file for a divorce soon,
but I'd like to prevent that if I can.


So I'm poking around on the downloads site under IDM looking for
patches, when I spot this:
IDM 3.5.1 Engine Patch2 20071213

And under fixes, I see this:
-Race condition in Remote Interface Shim connection code, where the
engine was not connecting back to the remote loader in cases where it
became disconnected.-

And I'm thinking to myself, "Hmm, I bet that's my problem right
there!", but before I go broadcasting an announcement to my
organization that the main file server is going offline for a reboot, I
wanted to know if others have seen an issue like this (it probably isn't
specific to just the AD remote loader) and whether that patch fixed it
for them.

I also wanted to know if I really have to unload DIRXML.NLM before
copying over the two replacement NLMs, because I just do not trust Java
father than I can throw the moon, and don't want it hanging my server up
again.

Anyways, any information that anyone can provide would be great. I
wish I didn't have to support a Windows server at all, but well, I'm
not omnipotent (though I sometimes claim to be), and thus can't banish
the things to the appropriate extra-planar dimension.


Thanks!,

--jkinard


--
jkinard
------------------------------------------------------------------------
jkinard's Profile: http://forums.novell.com/member.php?userid=993
View this thread: http://forums.novell.com/showthread.php?t=313764