Advice wanted on backup jobs (with containers)

[tom.starren]

Hi all,

Recently I've upgraded to a new Windows 2016 servers and attached new storage to this machine to act as ReFS repository in the future. Before I'm going to change all my jobs 1:1 to the new repository I want to take a look at my current job setup. Currently I've got 19 jobs for about 90 VMs, these jobs are categorized depending on function (domain controllers, web servers, etc), the need for application aware processing and VM size (to not exceed 1TB of vbk file). Now I'm thinking to move these smarter together and consolidate jobs and place them on the new ReFS repository. To get some inpiration, could some people share there setup (with folders, tags, etc)?

Thank you,
Tom

[Gostev]

Hi, Tom - with less than a hundred of VMs, I would create a single job by adding some highest-level container like vCenter, and enable per-VM backup files on the repository. This takes away a lot of complexity (and potential errors of missing some VMs), and will meet your VBK size requirement with zero on-going overhead of babysitting jobs as new VMs are added into the environment. Thanks!

[tom.starren]

Hello Anton,

Thank you for your information. I need more then 1 job because I have some machines which use application aware processing and some of them are backing up transaction logs each hour. Some other machines need to have a different policy regarding backup copies (but I could create separate backup copy jobs for this).

When using one large backup job would you pull everything out of the same storage snapshot or would you tweak this? When holding a storage snapshot for a "longer" time this could potential lead to a large snapshot.

[veremin]

I need more then 1 job because I have some machines which use application aware processing and some of them are backing up transaction logs each hour. Some other machines need to have a different policy regarding backup copies (but I could create separate backup copy jobs for this).

But this still can be achieved within one job by enabling/disabling application processing for particular VMs.

When using one large backup job would you pull everything out of the same storage snapshot or would you tweak this?

By storage snapshot you refer to one of the storage devices we have a special integration with?

[tom.starren]

But this still can be achieved within one job by enabling/disabling application processing for particular VMs.

Totally forgot about this feature, this could definitely work. Not all my SQL servers have the same guest prosessing features, so I would need a seperate job for this exclusion then?

By storage snapshot you refer to one of the storage devices we have a special integration with?

Yes we are using the NetApp storage integration.

[veremin]

Not all my SQL servers have the same guest prosessing features, so I would need a seperate job for this exclusion then?

No, you won't need a separate job. Either set "copy only" option or disable application processing completely for specific VMs (Guest Processing -> Applications -> Edit -> General).

[mma]

No, you won't need a separate job. Either set "copy only" option or disable application processing completely for specific VMs (Guest Processing -> Applications -> Edit -> General).

Make yourself happy and use vSphere tags for this.
You can use tags for credentails, Indexing, application processing, copy job, VM selection.....
It will take some time to create and assign the tags, but it's worth the extra work.

You will get a clean and dynamic job setup

[tom.starren]

So create a tag for a specific RPO/RTO and then seperate tags for application aware processing, indexing, etc.?

[foggy]

Yes, you can combine tags on a VM and that is their advantage over containers. Though please be aware of this issue.