Backing up from NetApp SnapMirror case#02454629

[resqguy]

I'm having trouble setting up a job to backup from a remote (DR) NetApp using SnapMirror. I have read the documentation several times and I still can't get it to work right. I have all the requirements called out in the documentation.

My Veeam B&R server is running at my primary location. I have a proxy and local repository at primary as well. This is used for short term restores.
I have a proxy at my DR site along with a repository. This is used for long retention of jobs

My NetApp at my primary site replicates to my NetApp at my DR using SnapMirror. This occupies most of my bandwidth between the sites.

Let me get to where I am unsure about the job config.

1) Source VM - do I specify the VM at the primary site (where it is powered on) or do I use the VM at the DR site (where it would be powered off)?
2) Proxy - I'm reasonable sure that I use the DR proxy, which is also where the target repository is located. (Y or N)
3) Target - I selected the DR repository which is where I want the job to land. - seems straight forward
4) Configure secondary destination ... is checked. Backup from Storage Snapshots is selected. - also straight forword
5) Secondary Target = NetApp SnapMirror as source. (this just seems strange to configure a source as a target)

That's it. When I specify the remote proxy it fails if I don't allow it to fail over to primary storage snapshot. Then it copies from the primary site instead of the DR.

What am I missing?

[foggy]

Hi Richard, basically, if everything is configured according to the documentation, then please work with our support engineer who will be able to define the reason of the failure to use the target storage snapshot after reviewing the job logs.

Just a couple of checks that are on the surface, so most likely you've already gone through them: both storages are added to Veeam B&R backup infrastructure and target proxy is properly configured and has a direct connection to the secondary storage array.

[resqguy]

Thanks for your response. I haven't had much luck in the past with Veeam Support. I'm sure there are good engineers there but up to this point I have had to solve my issues on my own.

I ran a test job by specifying the VM on the DR storage cluster as the source. It moves data but the B&R server does not track the stats. I'm also concerned that that backup would not be "application aware" since it would just be a copy of a NetApp SnapShot without Veeam managing it. When I specify the VM at the primary site I get an error Cannot find any relevant "NetApp SnapVault and SnapMirror relationships". I did a manual update but it still fails. Hopefully Veeam Support can help.

[foggy]

You can always use the "Talk to a Manager" option to provide your feedback, if you feel the case is not treated properly.

You should specify the VM in the production site, since it is the one you're going to backup. Please ensure volume SnapMirror relationship between the primary and secondary storage arrays is configured.

[resqguy]

Please ensure volume SnapMirror relationship between the primary and secondary storage arrays is configured.

I have been running SnapMirror successfully since the NetApp was installed.

Does the VM need to be the only one on the volume? Is there something to the specific configuration of the relationship that Veeam needs?

Here is a snip from my daily SnapMirror report:

SnapMirror History
SourceVolume TransferSize(GB) Replication Start TimeReplication End TimeTransferTimeTransfer Result
Corp_Temp_Mail 0.081 1/9/2018 0:10:00 1/9/2018 0:10:14 00:00:14 success
test_vol_ssd 12.869 1/9/2018 7:00:00 1/9/2018 7:07:47 00:07:47 success
VM_ACCT_FP 6.669 1/8/2018 10:05:00 1/8/2018 10:07:03 00:02:03 success

[foggy]

Does the VM need to be the only one on the volume?

No.

Is there something to the specific configuration of the relationship that Veeam needs?

None that I'm aware of. There are some licensing limitations though that should be taken into account.

[resqguy]

NetApp licensing is fine, we bought the Cadillac packages from both companies.

Veeam requested that we update to ONTAP 9.3, which took several months since this is a production environment. I have the same issue after the update. Opened case # 03018556 to continue troubleshooting. Now Veeam Support wants me to engage NetApp Support because Veeam "cannot find the storage snapshot for transfer". They think it might be a sync or permissions issue. The NetApp is showing the SnapMirror relationship as healthy and there are no errors.

I'm at a loss with what to ask NetApp to fix. Does anyone have this working?

[larry]

from a old post.
My issue when I first tried was I didn't make a mount point to the snap mirror. Open NetApp and create a mount on the protected volume just like it was a real volume. Then assign the rights to the mount point. Then when you browse in Veeam it looks like the snapshots at the main site. The really cool part is the instant restore, mount a VM from the mirror snapshot at the DR site in Veeam, this takes only a minute and the VM is running. My DR backups use this mount now which makes remote backups real fast because they are really local.

I also found I can have Veeam make local NetApp snapshots every hour ( doing for Exchange ) then I have Veeam tell NetApp to Snap Mirror all the days snaphots at night to the DR site. This allows me to send some snap mirrors every hour for the required RPO of some VMs while others snap shot only locall all day but then sends the days worth of snap shots to the DR site off hours.

At my DR site now I have Veeam backups but no Replicas, any more. I don't need the Replicas because Veeam can mount the VM from the snap mirror very fast and you then just vmonition or run it there. My RTO is very quick now.

One other thing I found out is to be sure to use a local server for FLR. If using a snapmirror at the DR site be sure to use the ESXi server at the DR site, it defaults to one at the other side, under advance you can chose the host to recover from.

The combo of NetApp and Veeam is powerful and simple once you get it setup.

[resqguy]

One other thing I found out is to be sure to use a local server for FLR

I was following your post until that statement. I'm curious what a local server has to do with a Female-Led Relationship.

I have more concerns after reading this https://forums.veeam.com/veeam-backup-r ... 48050.html. My boss isn't happy since Veeam Support specifically recommended ONTAP 9.3 to resolve this. If Veeam now says they won't have full support of 9.3 until later this year, they might get thrown out.

[larry]

my veeam is 9.5.0.1038 and netapp 8.3.2p6. The only data sent over the WAN is by NetApp, but veeam controls the jobs. I even have Veeam control the Netapp snapshots to SnapMirror only jobs. My jobs are all datastores sourced. I use a tags on the VMs is I want to skip.

The combo of NetApp and Veeam is still the best upgrade for backup and DR I ever did.

[foggy]

I was following your post until that statement. I'm curious what a local server has to do with a Female-Led Relationship.

File-Level Recovery

[orb]

One other thing I found out is to be sure to use a local server for FLR

I was following your post until that statement. I'm curious what a local server has to do with a Female-Led Relationship.

I have more concerns after reading this https://forums.veeam.com/veeam-backup-r ... 48050.html. My boss isn't happy since Veeam Support specifically recommended ONTAP 9.3 to resolve this. If Veeam now says they won't have full support of 9.3 until later this year, they might get thrown out.

I have been running this scenario on both 9.2 and 9.3 branch successfuly, even with cross replication between sites on a 8Mbit/s limited bandwith. What storage protocol do you use ?
I do run ONTAP9.3P5 without any trouble at all (when I configure it well )

Oli

[Andreas Neufert]

9.3 is supported:
NetApp FAS, FlexArray (V-Series), Edge VSA and IBM N Series (NetApp FAS OEM)
• NFS, Fibre Channel (FC) or iSCSI connectivity.
• Data ONTAP versions from 8.1 up to 9.3
• 7-mode or cluster-mode

[resqguy]

I wanted post a follow-up on this thread. I did get backup from SnapMirror to work but it took a lot of effort. The call got escalated to Tier 3 and they had to look at the code. The first thing wrong was the naming of the SVM/Vserver on the replica target NetApp. Our primary NetApp SVM was named svm1 and our target was named SVM1 (uppercase). I know, who does that? I wondered the same thing myself. I had to rename the target SVM. The next thing I had to do was change the account/user used for the remote/target Proxy server. We had just put in stronger security so only elevated user/accounts have remote access to servers.

I have jobs that will run this weekend. So far my test jobs ran file.

[foggy]

Hi Richard, thanks for the follow-up, yes, NetApp allows to have similarly named SVM/volumes, which doesn't play well with Veeam B&R (it is not case sensitive). We will mention this as not the best practice in our documentation.

[Adam.Bergh]

Just to follow up on this, our documentation and best practices will be updated to include a known issue of SVMs have identical names between the primary and secondary ONTAP systems.

[orb]

Does Veeam get confused as well if you give an alternate name during the relationship creation ? (forgot about the switch). Yon can always rename your svm with a different name or fqdn. You may have to update your snapmirror relationship.

Oli

[Andreas Neufert]

I think the issue is more, that we detect the second storage not correctly then.