Damn, I knew someone would ask for details and this is one I wasn't heavily
involved in. 

We were putting in W2K3 schema and some our company specific stuff. There
was something that collided with the E2K stuff - I want to say inetorgperson
though it was like many months ago and Exchange has killed too many brain
cells...

Ok, google helped... Here it is 

http://support.microsoft.com/default.aspx?kbid=314649


The interesting thing with this one was that we didn't see any errors on the
machine that we did the update on. Once it started replicating errors popped
on all the other DCs concerning the mangling. Once I caught wind of it our
onsite MCS guy and PSS guy chased it down and got me an LDIF file to correct
it. The article above basically. We applied that and everything was fine. 

I don't recall why it wasn't initially caught in the test lab but they
retried it and it did do it there as well. Most likely it was never
doublechecked on any machines outside of the one it was done on and a
comprehensive schema compare was never done. I can't recall though as I
really wasn't all over it as I had something else I was on and I can't even
remember what that was but it was critically important at the time.


As for the snapshot stuff...

Do an LDIF export of your schema or adfind dump or whatever. Make your
changes. Get another dump/export. Run Windiff against them and it will
highlight the differences. If someone really wanted to get fancy they would
script it (my preference would be perl) and have it output a delta LDIF
file. You could even do this against dissimilar forests to see what was
different or to update one forest to mirror another. 

The method described above is one of my old fallbacks for trying to figure
out what the heck is going on in the directory for various changes. It is
how I figured out the whole subnet/site business to script it years ago
(this was before Robbie and Richard's great book that didn't sell so well as
it is over most admins heads). Basically I would dump the whole config
container, do the crap in the GUI I wanted to figure out, then dump it again
and windiff to see what got changed. Then duplicated it in a script. 

  joe


 

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of marcus
Sent: Friday, January 30, 2004 8:54 PM
To: [EMAIL PROTECTED]
Subject: RE: [ActiveDir] schema updates

Same goes.  This is a relatively new topic for me.

To answer a few questions, the configuration is an empty root domain and
three child domains.  The extensions we are looking at are for Windows 2003,
Exchange 2003, and SMS 2003.  :)

Now Joe you mentioned something regarding taking a snapshot of changes in
LDIF format.  How exactly does this get achieved?  Thanks!

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Fugleberg, David A
Sent: Friday, January 30, 2004 3:24 PM
To: [EMAIL PROTECTED]
Subject: RE: [ActiveDir] schema updates

Joe - care to elaborate on the error that didn't become obvious until it
replicated ?  I'm just curious what to watch for - maybe I'll add some steps
to my schema change testing process...
Dave

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] Behalf Of joe
Sent: Friday, January 30, 2004 9:07 AM
To: [EMAIL PROTECTED]
Subject: RE: [ActiveDir] schema updates


Darn Vendors!!! Of course we could always crutch this by creating a Schema
diff file, snap the schema, update the schema, diff it, generate the ldif
ourselves. Not recommending that to anyone but is something I have been
thinking about. 

While we are on the topic of schema updates, one other recommendation I have
is...

The idea of having a DC off on its own that you do the change on is a great
idea. The initial idea was off on its own off the network, then it was off
in its own site with real long period of time for the site link (not greater
than a week though or it is ignored). The next step in this idea/method is
to put multiple servers in that site. That way you make the changes and then
they go to a small group of a couple of servers and you watch for errors on
ALL of them. We had an error in our last update that didn't become obvious
until it replicated. 

Also my next schema update I think I will coax the schema partition
replication through the system specifically by hand to see if it goes any
faster via repadmin (or possibly joeware) forced connection/partition
replication. 

  joe
 

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Tony Murray
Sent: Friday, January 30, 2004 9:47 AM
To: [EMAIL PROTECTED]
Subject: RE: [ActiveDir] schema updates

I completely agree with you Joe.  I've been hassling vendors left, right and
centre to provide LDIF files for schema extensions.  Unfortunately, noone
appears to listen.  The most recent extensions I've tested have been from MS
(SMS 2003) and HP (Managed Objects), both of which fail to provide LDIFs.
If we can't get the big boys to provide them, what hope do we have with the
smaller vendors.

VMWare with its snapshot facility is great for testing schema updates.
(Except for when you click the Revert button just next to Snaphot by
accident - something I've done more than once!)

Tony

---------- Original Message ----------------------------------
Wrom: PWIGYOKSTTZRCLBDXRQBGJSNBOHM
Reply-To: [EMAIL PROTECTED]
Date:  Fri, 30 Jan 2004 09:24:21 -0500

I will debate this one... :op
 
First no one should put in anything they don't completely trust. I allowed
that to happen once and now I have a bunch of attributes/objects out there
that have nothing to do with anything and almost certainly won't be used
because the people driving it didn't have a clue what they were doing but
had enough weight (at the time) to force the issue. Now I have them as an
example to give me weight for any other people whom I don't trust.
 
Schema changes are one thing that should be worked and reworked and reworked
in the lab until you are sure of the outcome. Pulling a live copy of
production into a disconnected lab for this is invaluable and in my view an
absolute requirement. 
 
Unless you find something that is an absolute "it must go in this specific
order" because of issues you have seen, try to put them all together and not
have the schema cache update until it hits the end. 
 
If people give you programs to run instead of LDIF files, beat on the vendor
as that is very bad on their part - Hear me MS, W2K3 forest prep pissed me
right the [EMAIL PROTECTED] off. 
 
Anyway, take the LDIFs and pull the schema cache update entries out and tie
them all together and try it in the lab to make sure everything goes well.
 
Note this is an OUTSTANDING place to use Virtual machines. 
 
  joe
 
 

  _____  

Wrom: KHJYFMYXOEAIJJPHSCRTNHGSWZIDREXCAX
[mailto:[EMAIL PROTECTED] On Behalf Of Roger Seielstad
Sent: Friday, January 30, 2004 7:37 AM
To: '[EMAIL PROTECTED]'
Subject: RE: [ActiveDir] schema updates



To me it depends if you're stacking like or unlike schema updates. For
instance, with Exchange 2000 there are 2 sets of updates - the ADC and the
Exchange proper ones. I'd stack those any day.

Now - if you're talking custom schema stuff, or extensions from companies
you don't completely trust, then maybe staggering them makes more sense.

The real question is are schema updates queued into the regular replication
interval or does the schema update process itself force replication?

--------------------------------------------------------------
Roger D. Seielstad - MTS MCSE MS-MVP
Sr. Systems Administrator
Inovis Inc. 

>  -----Original Message-----
> Wrom: ZOWCONEUQZAAFXISHJEXXIMQZUIVOTQNQEMSFDULHPQQWOYIY
mailto:[EMAIL PROTECTED] 
> Sent: Friday, January 30, 2004 12:39 AM 
> To:   [EMAIL PROTECTED] 
> Subject:      [ActiveDir] schema updates 
> 
> Sorry if this has come up before (haven't seen it hit the list). 
> 
> We have some schema updates to do... but was curious whether stacking 
> them up back to back was a very good idea since they all require full 
> gc syncs.  We've tested the extensions individually in the lab, and 
> they all act fine... any comments on why you should or should not do 
> this?  :-)



List info   : http://www.activedir.org/mail_list.htm
List FAQ    : http://www.activedir.org/list_faq.htm
List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/

List info   : http://www.activedir.org/mail_list.htm
List FAQ    : http://www.activedir.org/list_faq.htm
List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/
List info   : http://www.activedir.org/mail_list.htm
List FAQ    : http://www.activedir.org/list_faq.htm
List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/

List info   : http://www.activedir.org/mail_list.htm
List FAQ    : http://www.activedir.org/list_faq.htm
List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/

List info   : http://www.activedir.org/mail_list.htm
List FAQ    : http://www.activedir.org/list_faq.htm
List archive: http://www.mail-archive.com/activedir%40mail.activedir.org/

Reply via email to