A customer has two IP Replicated VNX systems: VNXA is the Primary system; VNXB is the disaster recovery (DR) system. VNXA is configured with VDMs containing CIFS servers, has secmap cache disabled, and runs a primary Usermapper. VNXB is configured as a DR target system and runs a secondary Usermapper.
During failover testing, VNXA is failed over and shut down to simulate a site outage. The small number of Linux users that access the CIFS data via NFS find that they no longer have the same permissions as when accessing from the primary site. Additionally, the UID ownership of many files has changed. On investigation it is found that the secondary Usermapper database contains far fewer, and different, mappings compared to the primary.
Why would this be the case?