Centera Timeout??? DR Testing

[GeebMan]

Is there a timeout setting somewhere in EV 8.0 for accessing Centera storage? We have two Centeras at different locations replicating between the two. We have our EV servers pointing to the 4 primary IPs of the local Centera and have the 4 IPs for the remote Centera in the replica list.

We were doing DR testing and shutdown the network ports for the primary Centera. We were able to pull files back from the secondary but there is a considerable delay before it re-hydrates the files. It took 3 1/2 minutes to copy back five 10K files. It appears that the EV server is trying each one of the IP addresses of the primary Centera, which takes about 30 seconds total, before it goes to the secondary Centera and rehydrates the file.

Is this normal behavior for EV to retry the primary over and over for every file that it is going to rehydrate? Is there a setting I am missing for removing this retry?

Thanks,

MC

[Liam_Finn1]

I dont think you are missing any settings. Centera will retry the primary and only default to the target once attempts have failed on the Primary.

[GeebMan]

So in a true DR scenario to get the performance we need we would have to take manual intervention and point the EV servers to the secondary Centera?

[Liam_Finn1]

No because if you do that then the data will be written to the destination centera and in that case you will need synchronous replication between the centera's to get the data back to the source if you change back

The normal DR scenario is that in the case of failure you get read only access to the DR or destination Centera. then when the Target or Source comes back online then and only then are writes made to the source Centera.

You have the DR scenario setup correct as far as I know. I dont think there is a way to force it to move to the destination and not try to query the source in the case of a DR issue with the source