Disk Duplicates and flip flopping CIs

[benjaminayo]

Hello,

This may be a simple fix but I am new to CMDB (1st month in role and CMDB coming from ITSM background) I am looking to resolve flip flopping and duplicate tasks being triggered on the storage disk table. When looking into the IRE issue I have identification rule set up for matching attributes = device id, name... 

Came to work this morning with new dup task with same name but different device id attributes. 

The other issue is midservers updating the attributes via discovery changing the name, device id, LUN...

looking at the HISTORY>LIST, seems to be the same mid servers changing these fields back and forth

Pease advise

[Vikram Reddy]

Hey @benjaminayo,

This is an identification rule design issue: device id and name are too volatile to use as identifier entries, so every time discovery flips one of those values IRE reads it as a new CI and fires a dup task instead of updating the existing record.

1. Identifier Entries on the disk class: drop device id and name, lead with serial number, or a stable parent device plus LUN combination if the rule is dependent.
2. Reconciliation Rules: set explicit data source precedence per attribute so your MID Servers stop overwriting each other, which is exactly what History > List is showing you.
3. Once the identifier is stable, merge or retire the existing duplicates so IRE stops matching new discovery runs against stale records.

Is this cmdb_ci_disk or a custom disk class? That changes which attributes are actually safe to key off.

Thank you,
Vikram Karety
Octigo Solutions INC

[benjaminayo]

How can you test #3, that the identifier is stable. I have found the module "Identification Simulation" but not understanding how you use this to test multiple CIs at a time.