What looks unusual here is not necessarily the Junos behavior itself, but the ROA design.
It is quite uncommon to see two different origin ASNs simultaneously authorized for overlapping space like this:
-
1.7.228.0/22-24 → AS9583
-
1.7.229.0/24-24 → AS4755
Usually this happens during ASN migration, partial delegation, DDoS mitigation scenarios, or temporary ROA misconfiguration.
On my MX304, the route is currently still being validated as valid via AS9583:
---------------------------------------
> show route 1.7.229.0/24
1.7.229.0/24 *[BGP/170] 19:01:17
AS path: 64049 55836 9583 I, validation-state: valid
-
> show validation database record 1.7.229.0/24
Prefix Origin-AS Session State Mismatch
1.7.229.0/24-24 4755 -- valid
-
> show validation database record 1.7.229.0/22
Prefix Origin-AS Session State Mismatch
1.7.228.0/22-24 9583 -- valid
1.7.229.0/24-24 4755 -- valid
---------------------------------------
So, if your router is still marking the route as invalid while the covering AS9583 ROA is present in the validation database, it may be related to route validation refresh/reprocessing timing inside the router, a stale validation state, or some local policy/version-specific behavior.
A route refresh, RPKI session reset, or forcing BGP re-evaluation could help confirm whether the route becomes valid after reprocessing.
------------------------------
Guilherme Contino Santana
IP Network Engineer
Link Brasil Telecomunicações
------------------------------