A customer I worked with recently had a need to create a consistency group that traversed clusters that were managed by the same XMS. They then wanted to use this CG to create a snapshot scheduler. To me, it sounds like the better route for this is RecoverPoint. They are aware of that, but have an immediate need for this at the array level. What other options do I have here? I'm thinking it's either RecoverPoint or AppSync.
Does the customer have their application volumes spread across multiple XtremIO clusters? I am trying to understand the use case of creating a CG across multiple clusters.
We don't support CGs across different XtremIO clusters (at the array level).
Sorry for the delayed reply. Yes they do. In fact, it's the OS that is mirrored at the host level between the two XtremIO clusters. If they take snapshots on only one XtremIO array, then they will have to follow a manual process to break the mirror, boot from the restored volume, and then re-create the mirror. I really believe they need to look at an alternative to their host based mirroring solution and possibly look at RecoverPoint to handle this relationship they are trying to create. If they have concerns about HA, then what I can suggest is half the hosts see XtremIO cluster 1 and the other half see cluster 2 and the two replicate between each other.
In the screenshot I attached above, M1 is the source and M2 is the mirror. The host is zoned to both XtremIO clusters and requires the pair to boot successfully unless they specify the volume to boot from.