migrate with HCX: basic questions

[Luca82]

Hi everyone, I'm trying to understand what exactly HCX does...I mean... it seems that with HCX I can do a lot of thing and this make me confusing. I'll start asking you a couple of questions. As far as I've understood, HCX leverages on vsphere replication engine (as well as SRM does) for DR and vmotion for migration. Thus I can replicate (for DR purpose) and move (for migration purpose). Is that correct?

Other questions:

(i) after a DR the original/production vms still remain on production environment....I mean, there are double vms (production site and a copy on DR site)...is it correct?

(ii) if otherwise I do a migration from A to B (hot/cold migration) the vm is moved therefore after the migration operation there are no vm on the A site, correct?

(iii) is it always need to implement NSX on target site?

[Enter123]

Hi,

I have used HCX to migrate cc 2000 VMs live from NSX-V to NSX-T. So basically network extensions and live vmotion. 

Now using it again to migrate from one VCF deployment to another: NSX-T to NSX-T, again network extension and live vmotion.

With this kind of migration there are no VMs on the source site after getting moved to the destination. 

I haven't used it in any other way. But what I have seen and experienced, it was pretty cool and I liked the product. 

[Sasidhar1234]

May I know how would you switch the gateways when you plan to evacuate the datacenter at the source? Does it requires downtime for changing the gateway at the Target? I mean at any point, if the Network extension with MON enabled from Source to Target has to be switched to only Target VXLAN. Please provide the steps and confirm whether it requires any downtime or not.

[Enter123]

After you migrate all Vms from one NSX-T segment to the destination, you have an option to unextend that network segment and click on a checkbox: "Connect cloud network to cloud edge gateway after unextending". This will make your destination network segment connected - but this will not disconnect automatically nsx-t network segment at source. This part I do manually and usually see 3 pings lost before the destination gateway becomes available. 

That is downtime which my environment was able to tolerate.