Copying the DB over could be a good first start.  Just make sure that you don’t 
have any changes in development that you don’t want to lose, because you will 
then have a perfect copy of production (data, forms and workflow), not just a 
copy of production data.  If you’ve never done it before, there are several 
threads on this list that discussed that topic and could be a good reference 
for you.

Just out of curiosity, how does TOPO “push” data into the CMDB?

I think you can work with a copy of the production data to do your testing.  
The key is that you want a way to get back to your original state.  That is, 
you make your copy, but keep a back up of the data that you can restore after 
doing some testing.  When you test your recon jobs, it will change your data 
(setting recon IDs, updating BMC.ASSET, etc.).  If you find that you have 
problems, sometimes you might want to revert back to clean data.  In that case, 
just having the CI data to reimport would be good rather than doing a complete 
database restore.  You can run RE jobs to purge the current datasets, and then 
reimport the CI data.  There are various ways to do this.  If you have DBAs to 
help you, the easiest way to do it might be to do a dump of all the CMDB 
related tables that you care about and simply truncate and then restore those 
tables’ data from the dump.  That will take a bit of work to work out for the 
first time, but once figured out, you’re set.

If you have AIE, you really could configure a job (or set of jobs) to pull the 
CI data directly from production.  If you keep the job scheduled, then any data 
that gets pushed to production from TOPO would then get pulled into development 
via AIE, keeping your discovery data up to date with what’s in production.

If you’re decommissioning the CMDB so soon, why are you so worried about 
updating recon jobs?

Normally, what I would do is get at least a representative sample of current CI 
data in development, develop and test the jobs, and then redo the same changes 
in our stage/test environment (which was a recently copy of production).  That 
verifies more fully that things are working correctly.  Once satisfied that 
things work correctly in stage, I would then update the recon jobs in 
production and ensure that they were working correctly there.

As someone else said, it’s easy to screw things up with recon jobs – the 
concept is great; the implementation could be improved greatly.  Do a lot of 
testing, and verify what is being matched and what is NOT being matched to make 
sure that you’re not missing anything, and you are including what you want.

Good luck,
Lyle

From: Action Request System discussion list(ARSList) 
[mailto:[email protected]] On Behalf Of Kathy Morris
Sent: Tuesday, March 23, 2010 10:28 AM
To: [email protected]
Subject: Re: RE Jobs

**
Will copying production CMDB over to development resolve this issue? Discovery 
is not pushing data to a development server.  Discovery pushes data directly to 
the production server TOPO.  I am not sure how I am going to be able to verify 
the RE jobs work if the data is different between development and production.  
Someone on my team is going to copy production to dev.  I thought to you need 
the cmdbdriver.exe to copy a dataset over to another server.  He's planning to 
just copy the database from production to dev.  Will copying the database alone 
work?

To test RE jobs, what will developers normally do? We are decommissioning our 
CMDB within the next few months so I have been testing RE jobs against datasets.
I would highly recommend building a dev environment.  Setting up reconciliation 
jobs is not necessarily an easy task and is very error prone.

Also, depending on how Discovery gets data into your production database, you 
may be able to have it feed both environments.  If Remedy is pulling data 
(using AIE to transfer the data) into production, you can simply set up the 
same AIE jobs on your dev server for it to copy the same data into that 
environment.  If, however, Discovery is pushing data into Remedy (less likely) 
then you’d have to look for other options.  In fact, one other option might be 
to configure AIE to actually pull CI data from production…

Lyle

From: Action Request System discussion list(ARSList) 
[mailto:[email protected]] On Behalf Of Kathy Morris
Sent: Friday, March 19, 2010 3:56 PM
To: [email protected]
Subject: RE Jobs

**
Hi All,

We do not have a CMDB dev environment.  I need to create RE jobs.  Is it better 
to copy BMC.ASSET into separate dataset, and run the RE job against the dataset 
copy to testing purposes.  Or should we build a separate DEV environment and 
create the RE job?  We also have only 1 Discovery Server that is connected to 
Production.  So the CMDB development server would not receive any asset updates 
from the Discovery tool.




_attend WWRUG10 www.wwrug.com ARSlist: "Where the Answers Are"_


NOTICE: This email message is for the sole use of the intended recipient(s) and 
may contain confidential and privileged information. Any unauthorized review, 
use, disclosure or distribution is prohibited. If you are not the intended 
recipient, please contact the sender by reply email and destroy all copies of 
the original message.

_attend WWRUG10 www.wwrug.com ARSlist: "Where the Answers Are"_
_attend WWRUG10 www.wwrug.com ARSlist: "Where the Answers Are"_

Reply via email to