Hi team,

Initially, my company started the replication with the SRDF/S mode between two VMAXs. The distance between two arrays is less than 30 km s. we use 2- 1 Gb link with twice the compression of data between 2 ciena 4200 boxes.

In order to test the DR site, we split the replication and activated the DR site to see the functionality.(We choose the time and days where there will be very less I/Os), Every thing was working well.

Later after 2 days, we re-established the replication in Adaptive disc copy mode in order to move the big amount of invalid tracks (data). We activated each device group at every half an hour time interval (we have 10 RDF device groups).

At this time we got a lot of latency on the host and the write service time was almost triple digit. All the data bases got hanged. I don’t know why and what went wrong.

I suspect, keeping in split mode for two days would have gathered lot of invalid tracks and activating them with in an interval of every half hr was a bad idea.

other than this i suspect the configuration of the VMAXs also.

the configuration is as follows:

enginuity :5875

mirrored cache :57735 MB

engines : 2

RA ports: 7h:0 and 10h:0

both the ports have the negotiation speed of 2Gb

these ports are connected to ciena 4200 boxes.

thin devices in the group has the skew of 65535 but the device group skew number is 32727. will this effect?

few of the meta devices in the group are concatenated and striped.

guys please help me out with your suggestion and let me know what went wrong and what will be a good approach. It will be great if you can support your answer with any references.

thank you