This post is more than 5 years old
4 Posts
0
13326
March 10th, 2015 10:00
VRTX - Server/Fabric/PERC/Virtual Disk - Understanding component failures
I'm trying to better understand the internal workings of the VRTX. Here is my setup to aid in my question:
Hyper-V 2012 failover cluster with 1 M520 as DC and 3 M620s as host servers with two mezz cards for fabric B and C.
Dual SPERC controllers (CSV and Witness virtual disks controlled by SPERC 1 and SPERC 2 as failover/backup)
I notice that the SPERCs are each connected to a fabric switch (SPERC 1 --> Fabric B Switch and SPERC 2 --> Fabric C switch).
What would happen if I were to have a fabric switch failure? Lets say its Fabric B switch, so I lose access to SPERC 1 and, consequently, my virtual disks. Since nothing is wrong with SPERC 1, will my virtual disks failover to SPERC 2?
And if anyone knows Hyper-V well enough, say the PCIe Mezz B card fails in one of my M620s, will the VMs on that host failover because they can't reach the CSV anymore?
Any thoughts are greatly appreciated!
Matt K
DELL-Chris H
7 Practitioner
•
9.7K Posts
•
48K Points
0
March 10th, 2015 14:00
Mattkane85,
With the Shared Perc 8's installed, and configured for Fault Tolerance, then it will function as you stated. The Perc cards are configured in a Active/Passive configuration, so when one Perc fails, the other Perc card takes control, which covers the controller. Now with your question, you asked that in the case when a switch fails will it still failover? That also would be a yes. With Multipath Support enabled, which is required with the Fault Tolerant Shared Perc 8 configuration, then it will use redundant physical components such as adapters, cables, and switches to create logical paths between the server and the storage device. In the event that one of these components fails, multipathing logic uses the alternate controller so that applications can still access their data.
You can find further information and details on these configurations here, on page 6.
Let me know if this helps.