SUMMARY: Warning for SunRay 2.0 CAM users

G'day folks,

Well, here's a summary without an initial question being posted.  It's a
cautionary tale of a thought-to-have-been-completely-successful SunRay Server
2.0 installation...

Several weeks ago, we had finished (or so we thought) a complex CAM setup
for a 1.2-2.0 upgrade.  Actually, we started with a "clean" system by just
dropping the 1.2 setup as there was no need to preserve anything.  Part of
the new configuration included a main "application" (using CAM's
nomenclature) script which read a number of configuration files and ran a
customised screen-saver and browser of choice on each kiosk (depending on the
MAC address of each terminal).  The browser (actually, a "starter script") is
run in a typical kiosk chroot'd environment.  All was tested and found to be
working perfectly.  So what was the problem?  Read on...

What we failed to do was reboot the server.  By using terminal specific
configurations and a controller script to call distinct browser-starter
scripts, we were able to test various configuration changes simply by
restarting individual sessions or the server software.  By the way, this
also provides an easy way around the "all your kiosks must run the same
application(s)" limitation described in the documentation.  Mind you, our
test procedure did give the client's responsible staff the opportunity to
learn the SRSS 2.0 management GUI.

A fortnight ago, we received a call saying that the server had been
specifically reset to correct another problem (which wasn't with the server,
but this turned out to be a good thing as it highlighted the troubles I am
about to describe) and that the DTUs were all showing just a "blank screen
with a big X".  I recognised this as the X server's built in screen-saver.

Luckily, we had retained the previous configuration on a second drive in the
server (the total job involves three servers and they aren't going to
officially switch over until the last has been done) and that was enabled
until we could get a look at the system out-of-hours.

Anyhow, sure enough none of the CAM programs seemed to be working and, at
first, I couldn't see why.  Literally, no changes had been made to the system
by anybody - anybody human that is.  It turns out that one of the scripts in
/etc/init.d does a bit of cleaning up as it installs a few libraries (among
other things).  The offending script is /etc/init.d/bbinit (part of the
SUNWbb package) and the offending code is a function called copy_dir.  That
function quite happily recursively deletes all contents of a target directory
and one of the calls to that function sets /usr/lib as the target.  The
result of this is that anything which has been specifically set up to allow
custom applications to run in the chroot'd environment will be deleted by
this script.

There were a number of potential remedies:
	1.  Modify the script
	2.  Take advantage of a configuration file .../config/copy.list
	3.  Write another boot-script to undo the damage done.

We chose option 3 as either 1 or 2 could be overwritten by a subsequent
patch or software upgrade.  So, we wrote a simple script which "puts back"
the libraries (among other things) required by the various programs used in
the total solution.

<gripe>
My biggest complaint is not that the "blow everything away" approach was
taken by bbinit, but that it isn't documented anywhere that I could find.  I
am a real Sun bigot (I've been using their hard-/soft- ware for years and, in
general, really like it) but the SRSS 2.0 software and associated documents
have been a real disappointment.
</gripe>

Anyhow, I hope the above tale of woe will prevent others from getting caught
in the same trap.

Ciao.

-- 
-------------------------------------------------------+---------------------
Daniel Baldoni BAppSc, PGradDipCompSci                 |  Technical Director
require 'std/disclaimer.pl'                            |  LcdS Pty. Ltd.
-------------------------------------------------------+  856B Canning Hwy
Phone/FAX:  +61-8-9364-8171                            |  Applecross
Mobile:     041-888-9794                               |  WA 6153
URL:        http://www.lcds.com.au/                    |  Australia
-------------------------------------------------------+---------------------
"Any time there's something so ridiculous that no rational systems programmer
 would even consider trying it, they send for me."; paraphrased from "King Of
 The Murgos" by David Eddings.  (I'm not good, just crazy)
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers