Solved

Metro Availability on vSphere

  • 6 August 2015
  • 5 replies
  • 1225 views

Badge +10
Hello-
Had a question on Nutanix's Metro Availability RTO for vms I was hoping someone could clarify. If you have a vm configured in Metro Availability and you have a DR situation that takes down the primary site or you proactively fail over the vm, in the case of a planned data center migration what happens to the state of the vm during that transition period? It is my undestanding the vm will temporarily drop while it is brought back online at the secondary site. Is this true? If true, could someone clarify what actually happens to the vm? Does it power off, etc? The reason I care is we are looking to migrate vms to a new datacenter and if the vm temporarily goes offline it will dramatically change my migration strategy. I was hoping the process would be similiar to a VMware Metro streched cluster where I could just vmotion between sites with no downtime.

Thanks,
Erik
icon

Best answer by mmcghee 7 August 2015, 03:20

Hi Erik,



If you don't mind, could you check out the Metro Availability best practices guide? http://go.nutanix.com/metro-availability.html



It goes into the exact detail you're looking for regarding planned migration and unplanned failures. I'm the author, so if you have any questions after reading those sections, please post again and I'll answer. But in a nutshell, a planned migration, where you activate the container in the secondary cluster, does require downtime today (VM power cycle). The planned migration scenario should be changing for the better this calendar year, so please let me know if you need more info.



Thanks,

Mike
View original

This topic has been closed for comments

5 replies

Userlevel 7
Badge +35
Thanks for the question elarson - I'll get back to you with some details
Userlevel 4
Badge +18
Hi Erik,

If you don't mind, could you check out the Metro Availability best practices guide? http://go.nutanix.com/metro-availability.html

It goes into the exact detail you're looking for regarding planned migration and unplanned failures. I'm the author, so if you have any questions after reading those sections, please post again and I'll answer. But in a nutshell, a planned migration, where you activate the container in the secondary cluster, does require downtime today (VM power cycle). The planned migration scenario should be changing for the better this calendar year, so please let me know if you need more info.

Thanks,
Mike
Badge +10
Thanks Mike. I will download the document for specifics. Have a great weekend.
Badge +10
Hello-
Finally got MA setup on NOS 4.7. Have a vSphere cluster with 8 nodes. Four at site A. Four at site B. Active vms configured to run on Site A with MA setup replicating from the active container to the standby container at the secondary site. Question is around how data is written. I understand if my primary site goes down, in an unplanned manner, once I promote the Standby container HA will power on the vms at the secondary site. Where I am confused is around how data is written if there is not an outage, but I manually vmotion a vm from the active site to the passive site? Since vmotion won't work unless it sees storage on the target host does the vm actually start writing to the container at the secondary site, in which case the Nutanix MA would be out of synch, or does it still write back to the active site in which case I guess Nutanix is somehow tricking VMware.
Thanks
Badge +10
Just got a hold of the latest document for MA and found the following....
however, standby containers are not available for direct VM traffic. The system forwards all I/O targeted for a standby container to the active site.

So, it looks like Nutanix redirects the active I/O back to the active container. Should have read the document better the first time..thanks