Pure Storage Active Cluster Integration

[timmi2704]

Dear all,
we are using Pure Storage integration without any real issues for some time now.
As we are using an active-active storage cluster, I was wondering whether anyone knows how to do the integration of both arrays of the active-active cluster?
Is there any configuration that I need to do or would it simply be enough to add the second array to the storage infrastructure within B&R?

Background:
With having all backup jobs targeting only one of two arrays, we see high latencies during backup hours. I'd really like to balance the backup load on both storage arrays.

Thanks a lot.

[foggy]

Hi Timo, Purity ActiveCluster is supported by Veeam B&R meaning that Veeam B&R generally works in such configuration. You can add both nodes to Veeam B&R, however, there's no way to point the job to this or that particular node as Veeam B&R selects one of them randomly (volumes have similar SCSI Unique ID). Though this could allow spreading the backup load somehow, provided proxies have access to both nodes.

[johannesk]

Hi,

We are also trying to backup vm's residing on a datastore on active-active Pure cluster.

Are there any plans on to make it possible to select the source array for the backup (assuming anything is possible)?

Regards,
Jóhannes

[foggy]

Hi Jóhannes, if you need to stick to processing a specific node, you can try to configure proxy access accordingly.

[canonex]

Hi,
does anybody already have more information about this?

We also have some issues with the Pure Active Cluster.
Very difficult to plan the jobs which proxy and repository to use.
And often lots of network usage between different sites.

Best regards

[foggy]

Hi Steven, are you using the latest plug-in from Pure Storage? Version 2.0.47 fully supports ActiveCluster synchronous replication relationships and you can control the source side using the 'Preferred Arrays' setting on the storage:

Source/target side selection for synchronous replication relationships occurs based on the host mapping information from the arrays.
- If a volume only on one side of the relationship is mapped to a host, that side is considered the source.
- If volumes on both sides are mapped, the 'Preferred Arrays' setting in the storage management console's host settings is used. If that setting is not configured, the first array in alphabetical order is considered the source.

[canonex]

Hi again,

We use the latest Pure plugin.

When performing datastore (Pure volumes and POD) backups they fail with 'cannot find a suitable proxy' although there are two proxies per site.

Preferred array has been configured for every volume (or POD) but still issues at one site (and not the other).

When we use a local volume instead of a stretched one, it works perfectly but with the active cluster volumes it fails.

[foggy]

Hi Steven, if you're still experiencing this, please open a case with technical support for a closer look.

[canonex]

Thanks. We have an open case with Veeam and Pure Support.
Both are investigating it and looking for the best method or workaround.