Can you please point me to a document with _detailed_ description of networking requirements for Files for PD recovery to work, as well as _detailed _ description of how the networking is being handled during the failover?
The reason I’m asking is that I’m planning PD for Files, and I need to understand the networking side of that, i.e. how Storage and Client networking are being handled during the failover to the other cluster, where the networking is different (different datacentre). Do I pre-configure portgroups to connect to in advance, as well as the IP-s to assign?, and AOS/Files handle this automatically during the failover? Or is this a “wizard”-driven workflow where I specify portgroups and IPs? Or is there a requirement for stretched VLAN-s for both Client and Storage networks? I wasn’t able to find and deatiled information on that, so am askign for your help.
Thank you, Victor
Page 1 / 1
Hi Victor,
When you chose to Activate Files Server on a DR site you are prompted to put in IP address ranges, port groups and such which means that you do need to have that information ready and port groups configured, DNS and AD working at the DR before you activate the FS.
To enable Files DR you will need to pair sites as part of Data Protection first. For that, the two clusters must be able to communicate via their VIPs. How you choose to achieve (via VPN, stretched VLAN, MPLS, VPLS or even SSH tunnel (not recommended as a primary option)) we leave up to you to choose. There’s more than one way to skin the cat, right?
The overview of the process is, you add Remote Site in the Data Protection tab. Add Protection domain for Files (should be created during FS creation). If it is, edit it and add the remote site in there.
Arrange networking on the remote site for the FS activation. Activate. I’ll add some documents that I think could be useful below.
Let me know if this helps or if you have further questions.
Hi Alona,
Thanks for replying to my question.
It is clear how to initially _configure_ the protection, thank you. The question was more about what exactly happens _during_ the failover. From your link to “ACTIVATING A FILE SERVER”, I can see that the portgroup and IP addresses need to be entered for Client Network, and suppose the same for Storage Network. Does it mean that AOS will re-IP FSVMs’ network interfaces on strat-up? And if so, do I gather it right that in general Files/FSVMs may be recovered to a cluster/datacentre with a completely different networking, as long as it’s routed between the two environments for replication? I.e. VLAN stretching is not a requirement?
Thank you.
Hi Victor,
Ah, I see now where exactly your confusion lies. Yes, you are correct. When you think of a FS it’s a cluster of file server VMs consisting of several VMs (1) with NICs (2) and disks (3) communicating to each other, to the CVMs, users and other servers. Out of those three components two are dynamic: NIC configuration and the user data can change. FS VMs OS is static.
That means that as long as you have the disks data replicated and the NICs configuration information available you can re-create the FSVM cluster on the DR site.
And that’s what happens. The FS VMs do not exist on the DR until you hit the activation button. Then essentially FS VMs are created, cluster is formed, configured with the IP config you provide upon activation and the user data is presented. Once completed you have an FSVM cluster on the DR site.
Does that help?
Hi Alona,
Yes it does help, thank you for the explanations.
In case when the FSVMs are re-IP-ed, the DNS records will need to be updated in order to the server name in a UNC path resolves to the new IP address. Given the Files are configured to use the AD DNS, will the DNS records be updated automatically by AOS/Files on FSVM startup, or will the cluster administrator need to edit the DNS records manually?
Thank you.
Hi Victor,
The first URL I included goes over instructions on how to trigger activation. Once you click the button the window below appears for you fill out. If you mean update of the existing AD records to point to the new server, they would need to be updated manually.
Hi Alona,
Yes I saw that screenshot at your URL, thank you.
And thanks for the answer - the DNS update has to be done manually after the Activation with different IPs.
This was probably why our earlier implementation of a replicated Files instance was configured with Client Network on a stretched VLAN… Which still keeps our Networks team unhappy. Would be nice if Files had the final step with DNS update in the Activation workflow. :)
Would be nice if Files had the final step with DNS update in the Activation workflow. :)