Current best practice for two-site B&R using v8?

[pkelly_sts]

Until recently we've had 3 sites, with one of them having a very small local VM estate but we used that sites infrastructure as a DR destination for the other two. On that basis I installed Enterprise Manager at the DR site and "pulled" replicas to the DR site so that in the event of a DR we had everything available at the DR site to just spin things up without too much manual intervention.

However, using the above configuration we can't realise the benefit of backup copy-based replicas because the backup copy is "sent" from the source site so the DR site replication jobs can't take advantage of it. It's not such a big issue as we have a good 100Mb p2p link between the sites so until now has been fine.

The part of the business at the DR site has now been sold so we're migrating from 3 to 2 sites & planning on doing cross-site backups & replicas, but I'm trying to decide on the best model to use, i.e. keep a B&R server at each site (we still use tapes extensively) or merging licenses & have a single B&R instance covering both sites, but then I can't decide which site should host that B&R instance. We currently only have a 10Mb link between the sites so bandwidth is significantly reduced.

It seems we have more flexibility in v8 as we can now separate the B&R instance from the "tape server" instance at each site so a single instance is more feasible than it previously was. I can't remember whether or not we're able to backup a B&R VM, with B&R? I guess we could potentially use the Endpoint product for that instance, especially as it'll be a dedicated VM so nothing outside of Veeam catalogs would be changing anyway.

Any previous discussions available based on changes that have been enabled by v8? (We've only just got around to upgrading to v8).

Regards,

Paul

[Shestakov]

Hello Paul,
If you have both sites working as production + DR, the best protection strategy is using both backup and replication techniques. And yes, you can make replicas use data from files moved by backup copy job. In general, having Veeam B&R server responsible for replication residing in DR site is the best practice.
Here on the forums are several related discussions. You can start with this one.

I can't remember whether or not we're able to backup a B&R VM, with B&R?

Yes, you can backup it. You can also schedule configuration backups.

Any previous discussions available based on changes that have been enabled by v8? (We've only just got around to upgrading to v8).

Discussions are all around the forums. Looks like the easiest way is to check What`s new document. Thanks!

[pkelly_sts]

Yes, both sites are production + DR and I intend to use backup copies & replication for each site. However, having B&R server in the DR site managing replication means that server /can't/ take advantage of backup copy data for replication, unless I actually have the B&R server in site A fully responsible for backups & replication on site B, and the B&R server on Site B fully responsible for activities affecting site A, but this feels a little "clunky" to me, hence why I'm wondering if I should just consider having a single B&R server managing all backups & replication for BOTH sites (and configuring a "tape server" in each site), and then also protect that B&R with replication so that, should I lose the site hosting that server, then I can get ops up & running again pretty quickly.

Sensible approach or not?

[foggy]

However, having B&R server in the DR site managing replication means that server /can't/ take advantage of backup copy data for replication, unless I actually have the B&R server in site A fully responsible for backups & replication on site B, and the B&R server on Site B fully responsible for activities affecting site A

Paul, could you please elaborate on what are your reasoning behind this? I don't see why single Veeam B&R server is not able of managing both sites in your scenario.

[gingerdazza]

We have similar scenario. Really I am looking to just have 2 B&R servers, one in each DC. Each DC will backup local VMs, and then pull the replication from the other DC. This is necessary so that if you've lost DC1, you may have lost B&R1 too so any failover plan needs to be at the other site. Both B&R servers will then do backup copy jobs to each other to equate to 3-2-1 rule.

[foggy]

With Veeam B&R configuration backup in place, you can fire up new Veeam B&R server and perform failover pretty quickly. Your setup is reasonable, however when we talk about replication from backup (which Paul is all about), using two backup servers makes them unaware of backups created by each other, complicating replication significantly.

[pkelly_sts]

Paul, could you please elaborate on what are your reasoning behind this? I don't see why single Veeam B&R server is not able of managing both sites in your scenario.

Sure, what I meant (but didn't say clearly) was "having [the second B&R] server in the DR site managing replication....", i.e. as you mentioned in your last reply, multiple B&R servers not being aware of the whole picture.

I might be fussing over nothing... Given my stated scenario, Site A/Site B, and the only problem I'm perceiving being unable to swiftly bring up VMs in the DR location due to "pushed" (rather than pulled) replicas, as long as I have the most recent copy of the config backup, would a newly deployed B&R server (in the DR site) be equally as capable of performing a replica failover, fully understanding its environment, as the original server would have been? If so it's probably not much of an issue after all.

In fact I may even try spinning up a test VM & deploying B&R to it, then restoring a config & just see for myself what it looks like...

Only other thing I can think of that might have a minor impact on things is timing of config backups. If memory serves me correctly, I think you can only (at least via the GUI) run a config backup once per day? Would be really useful if one could run a config backup/copy (to DR site) soon after the last replication job has completed, to fully protect the scenario I've been discussing.

Paul

[PTide]

Hi,

Would be really useful if one could run a config backup/copy (to DR site) soon after the last replication job has completed, to fully protect the scenario I've been discussing.

You can do that with a scheduled powershell script. Please refer to this PS guide.

Thank you.

[pkelly_sts]

Ah, thanks for that pointer, I shall investigate!