DFS failover issue

[leaden rune]

Hi Team,

I was working on DR test from one site to another , it’s a active active vserver and we are doing volume level replication
One we broke the replication and stopped the source vserver cifs we have found that DFS failed great but once we brought up replication and source vserver cifs service , we did see a lot of users DFS path on their drive properties was showing the DR DFS path as active and Source DFS path as not active

In both cases pre and post DR - we do have running vserver CIFS services, as DR sites also have some RW volume so we can not shut down DR CIFS services

I know reboot can solve this but I’m wondering even I was running CIFS session close , CIFS session was getting recreated .
For DR volumes were showing me “-“ if I do check CIFS session on DR post fail back .

DFS setting are set both as active and primary DFS properties as global preferences

Thank you

[molten cedar]

I think DFS assumes all DFS paths can be written to at any time, and if you set one DFS path to the DR destination, those volumes are read-only which will not work. I would suggest using SVM DR with identity preserve, which will keep the IP address of the source system on failover and everything will just work as it should, even without setting multiple DFS paths

[leaden rune]

Thank you , as this is volume based Replication so I can not set up identity preserve true , and DR Vserver has many RW volumes

[tulip dirge]

The best way to do this would be to have both shares configured in DFS. When you fall over to a new site, rename the old site cifs share name. This causes DFS to see the DR path as offline and directs everyone to the online path. DFS uses active and inactive as their descriptions. Else you can run a DFS flush command on all the windows devices, takes administrative level access I believe.