vcenter vs standalone hosts

[arsprod]

I've been running all my replication and backups through vcenter. Because of some recent issues with the vcenter server I changed all the jobs to local hosts. Now that I've resolved the vcenter issues I'm wondering if it's better to leave them replicating/backing up directly to standalone hosts or move back to vcenter. Is there a performance difference? Advantages/disadvantages to vcenter?

[kjc3303]

If you move VM's between hosts you do not need to change Veeam jobs is one benefit of using vcenter

[Vitaliy S.]

The one and the most important one!

[arsprod]

thanks!

[omorin]

Bonjour,

My question is quite simple, can VBR still backup or replicate a VM if the vcenter is out ?

In other words, is it better to directly acces VM by the esx server address instead of vcenter address in case of vcenter indisponibility ?

[Vitaliy S.]

My question is quite simple, can VBR still backup or replicate a VM if the vcenter is out ?

If your backup jobs are configured to use vCenter Server connection, then it is not possible to continue running these jobs if your vCenter Server goes down.

In other words, is it better to directly acces VM by the esx server address instead of vcenter address in case of vcenter indisponibility ?

No, it's not. Please see further clarification on this above. If I were you, the first thing I would do is recover the vCenter Server, if it is not possible, then of course the only option you have is to add standalone hosts and run your backup jobs using direct connection to the ESX(i) hosts.

[dellock6]

Another great reason to keep vCenter in a virtual machine. In case of problems, repidly boot vCenter from Instant VM Recovery and restart every activity from here.

Luca.

[flavor4real]

It is the recommendation to utilize the VMWare vSphere when it comes to Veeam. Now, is it possible to add several ESX/ ESXi connections instead of utilizing vSphere?

The point is to eliminate the single point of failure "vCenter" in the event of a vCenter crash.

[foggy]

Of course, you can add standalone ESX/ESXi hosts to Veeam B&R console (and it is the way to go in some particular cases, for example, when backing up vCenter VM). However, in case of moving VM's between hosts you will need to readd them to the corresponding jobs as their moref IDs will be changed during the migration. On the other hand, vCenter is vMotion-aware and your jobs will continue to run normally if they point to vCenter.

[veremin]

Also, kindly don’t mix two approaches in one environment. Having ESX(i) host added to backup console as a standalone server and as a part of vCenter might result in some issues. Thanks.

[flavor4real]

I see. So the best practice remains to have a vCenter connection setup. If vCenter fails, then I simple setup the esx/esxi connection and restore the vCenter VM.

I just thought to have the vCenter connection for all jobs, and a additional esx/esxi connection which would utilized for one backup strategy only, the vCenter backup/restore.

[veremin]

Yep, as mentioned above, you can create an additional job, using ESX(i) host as a standalone server, in order to backup vCenter VM that resides on it.

Thanks.

[foggy]

I just thought to have the vCenter connection for all jobs, and a additional esx/esxi connection which would utilized for one backup strategy only, the vCenter backup/restore.

Here is the dedicated topic regarding backing up vCenter, just FYI: Backup Replicate vCenter VM?

[ttester]

Just a best practice question: Is it better to add all hosts under infrastructure or just the vcenter? I was thinking of using only the hosts as it would make the vcenter a single point of failure for all hosts.

I'm only using veeam to create replicas that get put on a NAS.

[veremin]

Hi, Jay,

Your post has been merged into existing discussion regarding similar issue.

Having VMs added to backup console via vCenter makes backup/replication jobs vMotion aware. So to say, if some of the VMs is moved to different host, the jobs will be able to track this change, keeping backing up/replicating it without any issue. This isn't the same with standalone hosts.

For more information, please, refer to posts provided above.

Thanks.

[KristiEva]

I was wondering if anyone knew what the real benefits seen, (specifcally for replication & failover/failback operations), are when using Veeam with VCenter Server as opposed to working directly with the ESXi hosts?

The reason I'm asking is because in my environment, (a production site on one network (10.3.1.x) and my DR site on another (10.4.1.x), there seems to be only disadvantages to running jobs through VCenter server appliance (VCSA). The biggest disadvantage that I've come across that has been covered a bit on these forums is that in the event of my production site being destroyed (or the VCSA being otherwise unreachable by Veeam) I would have to manually spin up my VCSA replica at the DR site.

That sounds easy enough, until I consider the fact that I would have to update its IP address to communicate on the DR site network which could potentially cause some other issues & configuration work to be done before it could be used for failing over VMs (for example updating VCServer services, disconnecting/reconnecting hosts, etc).



In my environment where I'm using Veeam strictly for replication and failover to a DR site, would it not make more sense to run my replication jobs directly on the Hosts? That way in the even of a disaster nothing would have to be done with the VCSA and I could begin to failover VM's immediately?

[foggy]

Kristin, please find some considerations in the topic above.

As to re-IP issue, if you have Veeam B&R server responsible for replication located in your DR site, you can execute failover/failback procedure for vCenter machine. During this operation vCenter can be re-IPed automatically.

Btw, here is another good topic discussing similar questions: Question about failover with replication

[skipzoid]

I have three hosts each running three VM's each managed by vsphere.

When I migrate a VM from one host to another it breaks the backup and I have to visit the job and re-select the VM from the different host.

The VM's are still part of veeam's inventory the host in which they where on has moved. Can veeam automatically detect that the VM has moved location within the inventory and update the backup job accordingly ?

thanks,

[Vitaliy S.]

If you have standalone hosts (not managed by a vCenter Server), then this behavior is expected and let me clarify why. Every VM is tracked by its unique ID in the backup job and this ID is assigned by a host this VM is running on. If you migrate a VM to another host, then it receives a new ID thus backup and replication jobs treat this VM as a new one. If you want to prevent this from happening, you need to start using the vCenter Server, add it to the backup console and configure your jobs with the vCenter Server connection as opposed to standalone hosts. Hope this helps!

[skipzoid]

If you move VM's between hosts you do not need to change Veeam jobs is one benefit of using vcenter

I'm finding that when I move a VM between hosts, it breaks the backup job. I'm using vcenter to migrate the VM's from one host to another but then having to visit the veeam backup job to re-add the VM to the job.

[Vitaliy S.]

Did you register this vCenter Server in the backup console and added VMs to jobs using the vCenter Server connection? This is a must if you want to preserve these IDs during the migration process.

[skipzoid]

Yes, that's the exact process I have followed. I always move VM's via the migrate function in vcenter and the only connection I have setup in veeam is to vcenter, not directly to the hosts themselves.

[Vitaliy S.]

In this case, what you're seeing is not expected - having a vCenter Server registered in the backup console allows preserving the VM ID after migration. Please contact our support team to validate your config to make sure everything is set up correctly.

[skipzoid]

I've just migrated a VM from one host to another via vcenter and run a backup and it failed when it couldn't find the VM.
Having now gone back into the backup infrastructure and re-scanned the vcenter managed server and then re-run the backup job it has succeeded.
My guess is veeam periodically re-scans the vcenter server on its own I must have run the backup job before the scan had happened or the scan isnt happening ? - how often does veeam rescan if it does at all?

[Vitaliy S.]

My guess is veeam periodically re-scans the vcenter server on its own I must have run the backup job before the scan had happened or the scan isnt happening ?

Manual rescan is not required, it should be happening automatically as soon as you start the job. Let's ask our support team to take a look at the logs.

[PetrM]

Hi Skipzoid,

My guess is veeam periodically re-scans the vcenter server on its own

Veeam B&R server runs Broker service which collects and caches the virtual infrastructure topology, jobs and tasks take the information about
infrastructure toloplogy anb objects hierarchy from the Broker service.
May be it was too short period of time between VM migration and start of the job and Broker didn't get a chance to update information about migrated VM in its cache.

Thanks!