How to comply the AVAILABILITY for the Always-on Enterprise

[aschalk]

Hey guys,

soon I want add some business critical machine into veeam backup.
This machine hosts an orcale database which is used from everywhere around the globe, so it's pretty important for us.
The log backup has to do be done at least every 15 minutes.
Otherwise there is a high risk of the archive logs to run full. On the other hand we want the RTO as low as possible.

Now to the problematic part. How can I assure that the responsible job is running under all costs.

For example:
I want to supply the B&R server with some critical updates (OS, Veeam, etc.). This means a downtime of several minutes with a high risk that backup SLAs will not be met.

We also have the desire to push everything that is related to this databse into high availability, the backup should be no exemption.

So it comes all down to the question: How can Veeam assure that the backup SLA will be met at all times?

[aschalk]

No idea/solution to this?

[aschalk]

I tried to research a bit yesterday but it seems veeam truly has no way to achieve high availability atm.
Correct me if I am wrong.

Will this be considered for version 10?

[foggy]

If you need high availability of the entire VBR server for the case of disaster, you can use native hypervisor protection (like VMware/Hyper-V HA). There's no true automated high availability solution for your specific case, when you just need to perform an update of the server that requires reboot, but you can think of temporary switching to a second standby Veeam B&R server with the help of configuration backup/restore functionality.

[subglo]

Hello,

I have question for You all guys about HA for Veeam Backup and replication server. Veeam support answered that for now there is no such thing and that's why i'm asking You all guys.

We have multiple VBR servers with different licences and different purposes. One server is for vCloud organization backups, other is for cloud connect backups etc. Our servers are in shared environments for multiple clients. In time our client count will grow and so will those backup/replication/restore jobs with different schedules. We want to guarantee that VBR service is uninterrupted.

We have separate VBR's and SQL server which hold databases for those VBR's. All our Veeam infrastructure sits on VMware and there we have DRS and HA. For SQL server I can create SQL server cluster and guarantee HA if one DB server fails. How can i guarantee HA for my VBR service side if it goes down and i don't want backups to interrupt? Now it is like - "If our VBR fails, then it fails." Ofcourse i have configuration backups for VBR server and i can easily restore, but this is not what i'm talking about.

Waiting for Your opinions/experience.

[foggy]

Currently the workarounds you've described are the ways to achieve some sort of HA for the backup server, since there are no built-in capabilities for that.

[Gostev]

VMware vSphere Fault Tolerance is one possible solution in case your backup server is a VM.

[ajeeshb]

Hi,

I was looking for High Availability options for Veeam Backup server. Raised a support case and the answer from the Technician was not positive . I received below suggestions:

1. Create a configuration backup to an offsite location and restore it on a new Veeam server
2. Veeam Replication (with a lot of limitations like, no other jobs can be run at the same time a replication job is active)

But I am actually looking for a zero/very less downtime options. Like a Veeam server running on a cluster. Many of the enterprise level backup applications provide that options. Please suggest.

[PTide]

Hi and welcome to the community.

But I am actually looking for a zero/very less downtime options.

Currently there are not conventional means that are supported, however you may want to try these options:

1. Deploy a separate VBR and keep it in a standby mode so you can use it to import configuration backup
2. Deploy a separate VBR and use it to replicate your primary one.
3. Use Veeam Agent for Windows to backup VBR server.
4. Build a two-node SQL cluster to host Veeam database (a little bit extreme option that requires some manual work)

Thanks

[foggy]

A couple of other options in the thread above.

[ajeeshb]

@Gostev... When you say "VMware vSphere Fault Tolerance is one possible solution in case your backup server is a VM", do you have any documents from Veeam on how to implement this for Veeam BR? Or any steps provided by Veeam would be of great help.

[Gostev]

Basically you just install B&R into the VM, and make this VM fault tolerant in vSphere.

[billcouper]

This is a very interesting topic.

The licensing model means we run three VBR servers One of them is for Cloud Connect and for me this is the one that I would like 'always on'.

For the OP the primary concern is a mission critical database server on premise. It seems the request is that VBR support a HA config in and of itself, with multiple VBR servers that know about each other and use heartbeat to provide some level of failover to meet short RPO.

Proxy server automatic selection, combined with SOBR extent failover when unavailable, already provides a level of availability to the other components that allows maintenance to be performed easily. Having HA on the VBR would be the next step for the always on enterprise.

This seems like a legit request that would provide benefit to both the OP and my own situations.
+1 for VBR HA

[oliverL]

Or is there any way to achieve this by replicating the veeam management server via veeam integrated tools?

The only solution I see so far is with fault tolerance from VMware but unfortunately there is a sever under 6.0

[foggy]

Hi Oliver, please review this thread for some considerations.

[Iain_Green]

Afternoon,

A couple of questions..

• Is the backup & replication server able to be clustered in update 4?

• If not are there any plans to support this?

If no to these, do you have any suggestions on the below?

Our B&R server is always active, there are backups, copy job, sure backups and occasionally replications taking place for internal and numerous customers.

Getting outages for updates (OS patches) is becoming extremely difficult, which leaves us is the horrid position of having vulnerabilities left open until we can arrange a forced outage!

[Ryan S]

perhaps just create a separate job to do backup/replicate the server itself?

[Iain_Green]

Bad practice to replicate the B&R server, causes issues with the SQL server connection.

[AlexB82]

This would definitely be a nice feature request. Having a secondary sync'd VBR server would be so helpful. As mentioned, it's almost impossible to find an acceptable window in which to patch the OS or perform other maintenance on the box, especially if something goes wrong.....

[foggy]

Hi Iain, there's no built-in cluster capability in Veeam B&R, please consider one of the mentioned in this thread. There shouldn't be any issues with SQL Server connection either, since Veeam B&R excludes its own database from VSS processing. Thanks!

[Iain_Green]

Hi Foggy,

I have read this post and see no real answer to my question, most are excessive extra steps.

Building another VM!
Have multiple B&R servers!
Replicating the B&R server!

SQL can be clustered, that's on use for poor deployment of not setting that up to be clustered.
Veeam sells availability, yet its own product cannot be made truly available out of the box.

Making the VM HA is not the issue for me, I want to patch my server and limit the downtime.

Thank you though.

[skrause]

I know this would probably be a fairly difficult to implement feature but it would be nice if there was a way to run Backup & Replication in an HA configuration.

In theory I would see this as functioning like SQL Always On (and in conjunction with it for the backend DB, obviously) where the servers are part of a Failover cluster and the active job management would be handled by whichever server was "primary".

It wouldn't necessarily need to be an automatic failover scenario, but being able to know that if my VBR server went down "with the ship" in one datacenter I could have everything back up fairly quickly without needing to restore a configuration database or turn on a replicated VM etc. It would also make scheduling of different types of jobs a little easier for people who run backup jobs from a "local" VBR server in the production site and replication jobs from a separate VBR server in the target site.

[HannesK]

Hello,
your request totally makes sense. But as you said, it's probably fairly difficult to implement.

You already mentioned the "turn on replicated VM" as one of the most common workarounds. Beside of that, customers use VMware HA or VMware fault tolerance (if 2-4 CPU cores are good enough).

Your feature request is counted as +1

Best regards,
Hannes

[skrause]

I am thinking more of cross-datacenter HA, not just inside a single cluster (we already use vSphere HA).

I am going to do some experimenting with using our SQL Always On cluster as the core database server for the Veeam components when we setup our new data-centers.

[csydas]

Hi all,

Excuse my dumb question, but....

But as you said, it's probably fairly difficult to implement.

Why exactly is this so hard? The technology for SQL high availability is already there -- what is the hold-up from the Veeam side making this difficult? I'm not doubting you, but I'd like a better answer than "it's tough" when the HA part of the equation seems to have been solved by the technology we already invest in to use Veeam.

What about Veeam makes HA difficult when SQL already is HA?

[wishr]

Hello gentlemen,

We hope you understand, that we cannot comment on future release plans, roadmaps or feature ETAs, however, all your votes for this FR have been noted by Veeam Product Management team that is always with you here, on the forums.

Thanks

[DominikM]

+1 for a cluster functionality for VBR

[Maxrebo]

Hi,
I`m wondering if this is possible or the best way to go about this.
The other day the VM (Vmware) which had our Veeam Backup manager installed crashed overnight and subsequently the backups didn't run, I was wondering if there`s a way of having the Veeam backup manager installed on another VM so to act as a secondary manager if connection was lost to the primary Veeam enterprise manager?

[foggy]

I've merged your post into an existing discussion regarding similar matter. The most frequent advice in such cases is using vSphere Fault Tolerance.

[glamic26]

Without wanting to extract too much information from future roadmap etc but is it possible that V10 might see a more stateless/highly available deployment type from Veeam a la Rubrik and Cohesity? Their deployments obviously run across multiple nodes and are unaffected if a single node is lost other than any data streams occurring at the time of loss are temporarily interrupted. Unless this is likely to come soon from Veeam all of the people like those above in these threads needing a true highly available data protection product for their highly available workloads are likely to start moving away. Surely Veeam could achieve something similar by turning the proxy role into a clustered VBR role as well?