VSAN - to many restore points

[andreas2012]

Hi,

We have a VSAN system with several hosts.
The VM`s off course move between them.
We have 2 vm proxy pr vmware host
We have configured a backup job pr vmware host
When we now run our backup, I can see that the VM`s have created many more restore points that i thought it should. Is that because when a VM move over to another vmware host, it does not "know" the other restore points?

As you can see from the images below, we have ESX22, ESX23 and ESX24 that this VM has moved between. The jobs are configured to keep 7 days.
ESX22 full synthetic Tuesday
ESX23 full synthetic Tuesday
ESX24 full synthetic Wednesday

https://ibb.co/jDzmcPK (ESX22)
https://ibb.co/TwzFptg (ESX23)
https://ibb.co/DbLYLNQ (ESX24)
https://ibb.co/MnqyBqt (Restore points)


What I would like to achieve is this: 1 full backup and the rest incremental. After 1 week it should start over.... So how should i fix this ?

Thanks for any reply.

/R
Andy

[Andreas Neufert]

Yes, your assumption is correct. If you backup per ESXi host, then each chain is independant and when a VM moves it is seen as new (if not processed already in this same job) and creates a new active full for it.

Why do you have structured the Job that way?

If you worry about the optimal path for the Proxy selection on same host as the VM runs, then this is automatically the case with HotAdd processing. We detect that there is vSAN and preferably use the proxies on the same host. Independant if you choose a single host as source.
https://helpcenter.veeam.com/docs/backu ... ml?ver=110
So you can potentially just have one job and the optimal pathes are selected.
Or you can structure the Jobs in another way that it makes sense for you (gold,silver,test,....)

[andreas2012]

Hi,

Thanks for the reply.
This setup has been there for some years, and I believe it was structured like this because we had to use proxies on the same host. So the "We detect that there is vSAN and preferably use the proxies on the same host" how long has this been there, if it always has been there, then I guess we have misunderstood.

I definitely don`t want to have it like it is today.....

So as you say I can create 1 job for "all" vms, that would solve my problem, because I guess there is no other way to solve it as long as the VM`s are floating around ?. In my mind that should not be a problem, the job will just take awhile. It is around 120 VMS, not sure about how many TB....

Comments ?

/R
Andy

[Andreas Neufert]

I would say the special vSAN processing is there since some years.
Your statement about the job is correct.

I would maybe split it a bit for flexibility reasons (think about you want to perform a new active full or change the guest processing credentials).
Smaller Jobs would give you more flexibility. If you worry about new VMs, then you can make one job on cluster level and exclude the VMs from the other Jobs.
That way you catch all new ones automatically while keeping the flexibility of multiple jobs for specific purposes.

[andreas2012]

Hi,

Right. Thank you for good information, and explanation.
I will change my job`s and if there is any other questions I will post them

/R
Andy

[andreas2012]

Hi again,

I created a new backup job, and selected individual VMs, but the system is not using "optimal path for the Proxy selection on same host as the VM runs".
I can see from the backup job, that vms are using hotadd, but the proxy that is used is not on the same ESXi host, it is actually not on the same location either.
For example

VM1 is located on ESXi host 1, and the host is located on first floor. But the proxy that is used is located on ESXi host 2 which is located in the 2 floor. Suggestions ?

https://ibb.co/L8dptVm

/R
Andy

[Andreas Neufert]

I think by default there is a 95% data amount best proxy path backed into the code as it will help with optimal timewindow processing.
Can you check, was it only one VM or is it not working at all?
Did you added the vcenter or only the ESXi hosts to Veeam?

[andreas2012]

Hi,

Its all the VM`s, so now we have much more traffic between the two floors.... That's not something we want
I have only added vCenter to Veeam.

Suggestions ?

/R
Andy

[Andreas Neufert]

This is not normal. Can you please open a support ticket and upload logs, so that the support team can have a look at it?

Did you had maybe selected specific Proxies in the Job or does it stand on "Auto"?
Are the requirement met for the Proxies that they have to run on the vSAN Datastore?
https://helpcenter.veeam.com/docs/backu ... ml?ver=110

[andreas2012]

Hi,

Ok, I have now created a support case.
Proxies are selected as Auto.
Proxies are located on vsan storage.

Thanks for help.

/R
Andy