I have a current situation with 2 NS-120s using Celerra Replicator migrating to 2 VNX’s using Replicator in a bidirectional configuration, they are using CIFS for production at both sites but have the primary Usermapper enabled for both data movers on the NS-120's.
I have detailed the configuration below:
2 NS-120’s in a Bi-Directional Celerra Replication configuration.
Site 1 has its own VDM/CIFS Server/File Systems being replicated to Site 2
Site 2 has its own VDM/CIFS Server/File Systems being replicated to Site 1
(Both are used for production, one domain).
Each also has its own home directories and no quotas are being used.
Both have server_2 as the primary Usermapper.
Luckily, they've never performed a failover and have avoided the consequences of duplicate UID/GID to SID Mappings.
They are migrating to two VNX's in Site 1 and Site 2.
I currently have vdm/file system replication sessions from:
Celerra Site 1--> VNX Site 1 -- > VNX Site 2
Celerra Site 2--> VNX Site 2 ---> VNX Site 1
for bi-directional replication.
I have set VNX Site 1 as the primary Usermapper and VNX Site 2 as secondary Usermapper
I have uploaded the UID/GID Usermapper Database from Celerra Site 1 and downloaded to VNX Site 1
I haven't done anything with the VNX Site 2 besides set it as a secondary Usermapper.
My initial plan was to do just a switchover from the Celerra Site 1 to VNX Site 1 and for Celerra Site 2 to VNX Site 2 was during the maintenance window, delete all secmap cache entries from the Celerra Site 2 and then perform a switchover to VNX Site 2.
My thinking was that once the final data transfer takes place for the VDM Replication it will remove the secmap from the replicated VDM in VNX Site 2 but was told users could still have problems accessing old files.
Any help or direction on the best way to resolve the dual primary usermapper configuraiton issue.