Software Installations: Active install = false for single installs

[pavan Punna]

 

Hi all,

We are observing several software installation records in cmdb_sam_sw_install where:

• Active install = false
• Deduplication processed = true
• The CI currently shows only one installation
• The installation is from a single discovery source
• There are no visible duplicates

In some cases (for example Windows Server–related installs), this occurs at noticeable scale (thousands of records), which raised questions for us around correctness, reporting, and remediation.

What we have learned so far

After working with ServiceNow HI and reviewing multiple scenarios, here is what we understand:

1. Active install = false does not always mean incorrect data

   • This can be expected behavior if:
     • Another installation (possibly from a different discovery source) existed historically and was selected as the primary install.
     • That historical record was later deleted or cleaned up.
   • The remaining install can stay inactive indefinitely.

2. Same-source installs are not deduplicated

   • The dedup job only deduplicates across different discovery sources.
   • Multiple installs from the same source are assumed to be handled by the source itself.

3. There is no out-of-box auto-correction

   • There does not appear to be any background job that automatically “fixes” inactive installs once deduplication has already occurred.

Open questions we would like community input on

1. Is it acceptable to leave these records as Active install = false?

   If an environment has a large number of software installation records marked Active install = false—including cases where only one installation exists on a CI from a single discovery source—is it acceptable and safe to leave these records as‑is?
   Will this have any impact on SAM reconciliation, licensing calculations, reporting accuracy, or performance, assuming reports rely on Active install = true?
   •  

2. For single‑install scenarios (single CI, single discovery source, no duplicates) where Deduplication processed = true and Active install = false, is there any recommended or supported way to:

   • Identify how many such records exist at scale, and
   • Correct them (for example, selectively resetting the deduplication flag and re‑running the dedup job)?
     How do others validate whether a historical primary install from another discovery source actually existed before deciding to remediate?