Newbie question: Hot add vs. network mode

[marius roma]

Where can I locate documentation about "hot add" mode compared to "network" mode?
I get warnings saying that "Hot add is not supported for this disk, failing over to network mode..." but I can't understand the real difference and the way to let disks support it.
Is the problem related to virtual disks (.vmdk) or NAS disks?
Regards
marius

[foggy]

Mario, all the information is available in our sticky FAQ topic.

[mshorts]

where is the sticky faq? it would be nice if you added a link

[Vitaliy S.]

It's really hard to miss it... see the sticky topic on top of this very forum, that is even highlighted with the red font

[neilmurphy65]

Hi all.

At the risk of asking a question that may have already been answered elsewhere, I'm hoping someone can clarify this for me.

Say I have a physical VBR server backing up VMs on a vSphere cluster. The VBR server is also the repository. The VBR has no direct SAN access so my choices would be Network (NBD) mode or virtual appliance (VA) mode. Best practice would suggest that VA mode is preferable. But what makes it better? What's different about the data path? Is there a difference in the number of concurrent processes?

Many thanks,
Neil.

[Shestakov]

Hi Neil,
In the Virtual appliance mode, VBR uses the VMware SCSI HotAdd capability that allows attaching devices to a VM while the VM is running. During backup, replication or restore disks of the processed VM are attached to the backup proxy. VM data is retrieved or written directly from/to the datastore, instead of going through the network.

[alanbolte]

I recommend consulting the new best practices guide:
https://bp.veeam.expert/resource_planni ... modes.html
https://bp.veeam.expert/resource_planni ... _mode.html
https://bp.veeam.expert/resource_planni ... _mode.html

[neilmurphy65]

Thanks Alan.

So, the best practice says "this mode supports a 100% virtual deployment using direct storage access through VMware ESXi storage I/O stack having very little overhead". I get that. What I'm trying to understand is, if the repository is out on the network, say a JBOD attached to a physical backup server or a NAS device, is there any benefit to using the virtual appliance method and,m if so, why. It seems to me that the data path will be the same. The reason I ask is that I teach Veeam presales courses and the quetsion has come up a couple of times.

Cheers,
Neil.

[foggy]

Repository location is a separate question, data retrieval method affects the way of getting the data from production datastore to the proxy server. The benefit hotadd provides comes from the fact that data is transferred to the proxy server via ESXi storage stack, which is much faster, instead of the network stack. Further, the data is sent (compressed) from proxy to the repository - over network, in the scenario described in your last post. If you want to avoid that, consider utilizing direct SAN right from the repository server (this, however, is not possible with local storage).

[neilmurphy65]

Thanks foggy. Now I understand - it's the difference between using the ESXi storage and networking stacks.

[Zew]

This was a great thread! ANd great information, I do however have a couple questions.

1) Has DirectNFS, and DirectSAN been merged now into Direct Storage Access Transport mode?

"Depending on the connection the proxy can be deployed as follows:
On a physical server with FibreChannel, FCoE, iSCSI or NFS
On a VM with iSCSI and NFS"

2) CBT is not used when Virtual Appliance Mode is used.
"Note: Changed block tracking (CBT) will be disabled for these hot-add proxies. Consider that it may impact the backup window in case the said virtual machines should be included in backup or replication jobs."
In my Case I use an All-in-One Veeam on a VM within my cluster. As of right now most of my VMs are in a NFS datastore shared by my hosts.
My Veeam VM has a leg in the NFS network via a added VMPG on my hosts networking stack, for this one VM allowing direct NFS to work (And works great). I recent started making changes to move over to
iSCSI so I could utilize MPIO and get faster single session activities done (svMotion in particular). I moved one of my VM's to the new iSCSI datastore. However checking the results of the Job I got this:

"Using Backup Proxy VMware backup proxy for Hard Disk 1 [hotadd]
Hard Disk 1 (300 GB) 8GB read @ 60 MB/s [CBT]"

I already have iSCSI IPs for my Veeam VM for it's local repo to another SAN (using the same 2 seperated iSCSI subnets) so at this point as long as I add the target IP of the production VMware to the Veeams windows iSCSI initiator settings, make sure I can see the disc in disk manager (don't initialize it, of course) and then I'm assuming it would change form hotadd to DirectSAN. But my questio is if CBT isn't supposedly used in Virtual Appliance mode, why does my Veeam backup job say CBT after saying it's using [hotadd] mode?

[foggy]

1) Has DirectNFS, and DirectSAN been merged now into Direct Storage Access Transport mode?

Yes, Direct Storage Access setting includes both Direct SAN and Direct NFS modes.

2) CBT is not used when Virtual Appliance Mode is used.
"Note: Changed block tracking (CBT) will be disabled for these hot-add proxies. Consider that it may impact the backup window in case the said virtual machines should be included in backup or replication jobs."

This relates to backing up hotadd proxy VMs themselves. CBT woks on other VMs fine.

[Zew]

This relates to backing up hotadd proxy VMs themselves. CBT woks on other VMs fine.

OHHHHHHHH, like if I was to do a backup job of my standalone Veeam server (Backing up itself)?

[Shestakov]

Yes, backup proxy to be exact.
And in default configuration backup server has a backup proxy role.