Description of problem:

The duplicate CVEs in the errata are caused by one CVE having 2 or more urls (See below). 
[#<Katello::ErratumCve:0x0000560e9b405038
id: 524614,
erratum_id: 497834,
cve_id: "CVE-2025-37803",
href: "
https://access.redhat.com/security/cve/CVE-2025-37803
">,
#<Katello::ErratumCve:0x0000560e9b35e620
id: 554209,
erratum_id: 497834,
cve_id: "CVE-2025-37803",
href: "
https://www.redhat.com/security/data/cve/CVE-2025-37803.html
">,
 

I am not exactly sure which url is expected. Both urls worked. The "www.redhat.com" one simply redirect to "access.redhat.com"

I guess this could be the cause. We synced "Repo A" that contained "erratum A" (with CVE href 'A') a few weeks ago. Later we sync "Repo B" that also contains "erratum A" (but with CVE href 'B').

How reproducible:

Not sure.  I failed to reproduce the issue even I had tried to sync some affected repos for the errata. It appears to be only reproducible before the CDN repo metadata is corrected.

Acceptance Criteria

1. Add some scope magic to deduplicate and pick the first/last href
2. Add some rake task to clean up if needed

Testing steps

1. Create 2 fixture yum repos with the same errata but different href (tweak updateinfo xml)
2. Verify that the errata is not duplicated.

Is this issue a regression from an earlier version:

I believe it is not a regression

Business Impact / Additional info:

To workaround the report issue, we can add a ".uniq" to the report template.

'CVEs': erratum.cves.map { |c| c.cve_id }.uniq,