I am working with a latin American account tean where their customer requires a non-disruptive HA failover solution. The customer is setting up a video streaming service similar to Netflix. They want their 3TB oracle 11g catalog to never have disruptions.
They will be purchasing two VNX5300s to be linked via FC within the same datacenter. They will have HP servers with VMware attached to both VNXs. They want to have non-disruptive high availability of their catalog in the event of a storage or server failure. They are inquiring about a single engine VPLEX solution. However, they believe that VPLEX Metro, with VPLEX Witness, is the correct solution?
Does VPLEX Local support non-disruptive failover? Or, is VPLEX Metro the preferred soliution?
Thank you Ankur. I have pointed the account team to your post. Your diagram shows 100GB but I suspect that a 3TB data load will not be an issue.
My environment has two VNX 5700 arrays behind a VPLEX local cluster.
I just conducted a crude failover test by removing a LUN from vplex view on the VNX storage group level (removed the LUN from the Storage Group on one of the VNX arrays).
There’s approx 17 seconds where the volume is not writable to – though the OS and the system does not throw any errors. We are testing by outputting a “date” command every second to a test file.
Could you let us know if the 17 seconds is normal behavior? Is there any way to tune that?