Cleaning CMDB Tables
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-20-2014 05:59 AM
I am interested to know if It is recommended to clean the CMDB up to start from scratch, following this plan:
1) TURN OFF discovery schedules.
2) Clean up CMDB except business services (because these are already in use).
3) Upload inventory manually from excel through transform maps.
2) TURN ON discovery schedules to refresh and update inventory with configuration information.
Any thoughts?
Thanks,
Jesus E. Melendez M.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-20-2014 06:11 AM
If the data is not in use then I don't think there's any issue with cleaning it out and starting again. Any time you delete data, the biggest problem you run into is that the data has already been referenced somewhere. As long as you're confident that the data isn't being used you should be safe to delete it and start over again. If it is being used, then you should find a way to coalesce on the existing data instead so that you maintain any system relationships and references.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-20-2014 06:13 AM
Hi Jesus,
I went through this same scenario a while back, and my only "tip" is to make sure you realize there are a lot of tables involved in a clean-up of the CMDB - especially if you have already used a tool such as Discovery. Things like network adapters, routing rules, etc are populated and will require you to delete if you truly want to clean things up and start from scratch. These are things that aren't usually user-facing as much as servers, routers, etc. and may easily be overlooked. Look at the CMDB Schema map and you will see the different table that may contain data.
I would also recommend NOT uploading data via XLS if you are going to turn Discovery back on. Just let the tool do the work. This will also reduce the chances of duplicate CI's or other bad data since you will solely be looking at data returned by Discovery and not a mixture of discovered data and manually entered data.
Hopefully this helps some. Good luck!
Jason
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-20-2014 06:16 AM
Jesus, I have a few questions:
- First, why are you considering such an extreme method to start over? Is the data inaccurate? Is it just stale?
- What version of ServiceNow are you on? I ask about this to make sure you define your asset management configuration with Model Categories before you get too far. It will be easier to work with if you have this in mind as you clean up.
- What is your source for the manual import? Is it really inventory, or is it items you expect to be in use?
- Why this other source instead of Discovery?
- I agree with leaving the Business Services, as those are not discovered anyway.
- Are any of your current CIs linked to, either in business service maps or just in Incidents, Problems, and Changes? That could have an impact.
- With this in mind, you may consider cloning your prod to dev and testing this in dev before doing anything in production to make sure the process works as you intend.
Sorry to "answer" your question with more questions...
Ben
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
08-20-2014 07:36 AM
- First, why are you considering such an extreme method to start over? Is the data inaccurate? Is it just stale?
Yes, too much data, data is not accurate, there is probably much more information of what the business need (overwhelming). I am anyways considering limiting the access by user roles and groups
- What version of ServiceNow are you on? I ask about this to make sure you define your asset management configuration with Model Categories before you get too far. It will be easier to work with if you have this in mind as you clean up.
Current version is Dublin.
- What is your source for the manual import? Is it really inventory, or is it items you expect to be in use?
These are inventories in Excel (accurate data). The unique key we use is the device name. If Imported into the CMDB, I need to ensure that when I turn ON the discovery — records will not be duplicated (I understand discovery will only update the records where the CI_name = CI_name and not duplicate.
- Why this other source instead of Discovery?
Due to SNMP and WMI limited configuration on the entire network, there are devices that are not being discovered. That's why the excel inventory source is still more reliable. (The SNMP is something we are working on, however I am wondering about this path to populate CMDB as a short term strategy).
- I agree with leaving the Business Services, as those are not discovered anyway.
- Correct.
- Are any of your current CIs linked to, either in business service maps or just in Incidents, Problems, and Changes? That could have an impact.
Only a few — maybe less than 100 records.
- With this in mind, you may consider cloning your prod to dev and testing this in dev before doing anything in production to make sure the process works as you intend.
Absolutely, I forgot to mention above, but if doing so, I will definitely do it in a testing instance.
