This post is more than 5 years old
1 Rookie
•
11 Posts
0
1614
need understanding on Vblock SSHPD
Below is the snippet pasted from EMC.Please help me understand SSHPD,XtremSW and XtremSW
the Vblock SSHPD uses XtremSF cards with FAST VP to accelerate reads at multiple levels and writes in the storage array. XtremSW writes data at the array LUN and flash device and upon acknowledgement will make the data available on the XtremSF card. Thus, both database reads and writes are accelerated with the Vblock SSHPD.
Best regards,
Vishal
jeff_browning
256 Posts
1
October 29th, 2014 10:00
Vishal:
To your questions:
Yes, that's correct. VPLEX supports 5 ms of additional latency. This provides approximately 100 KM of distance between the two sites, but that is very configuration-specific. The important limit is the latency, not the distance. For more information on this, see this wp:
White Paper H8930: Oracle Real Application Clusters (RAC) on Extended Distance Clusters with EMC VPLEX Metro Best Practi…
Yes, that is exactly how Oracle RAC with VPLEX stretched cluster works. There is a bit of additional latency in the interconnect due to the distance. Speed of light and all that.
Exactly what VPLEX does. So, yes, your RAC cluster stays up, and all traffic will now route to the surviving nodes of the clsuter.
Regards,
Jeff
jeff_browning
256 Posts
1
October 29th, 2014 04:00
Vishal:
Here are some items on the SSHPD:
Regards,
Jeff
DBA_Vishal
1 Rookie
1 Rookie
•
11 Posts
0
October 29th, 2014 08:00
Hi Jeff,
Thanks for sharing the info.
1)I understand that extended RAC would mean that extending the storage for RAC to another data center which can go upto 100 KM and possible using VPLEX .Please confirm my understanding.
2)Additionally,please confirm if (assuming 4 nodes in RAC cluster) the nodes can also be separated across data centers which would mean that 2 nodes in one data center and 2 nodes in the other I think that way private interconnect speed would go down.Please confirm if that is feasible.
3)Let's assume that my one data center goes down would the node automatically access data from the other data center without any downtime? Is there a replication which keeps on happening in extended RAC
Best regards,
Vishal