Our second release today which supports the new PowerMaxOS 10 and thus REST API 10, as well as the new platforms PowerMax 2500/8500, is Dell (nix EMC) Storage Analytics or ESA 7.0. I’m focusing on PowerMax obviously, but there are lots of other features for the other platforms if you use those.
The main features of the release for PowerMax are the following, which encompass the PowerMaxOS 10 support:
- PowerMax File
- PowerMax NVMe/TCP
There is also an enhancement to report the path and volume identifier for PowerMax devices.
As always, the compatibility matrix is on our e-Lab navigator here.
Before getting into the two new features, I want to explain something about the metrics with REST 10.
I want to call out the reported metrics since there have been significant changes with REST 10 compared to REST 9. First, you’ll notice all metrics have lost the “EMC” tag. I’ve covered the end of that name in a number of posts so it should be of no surprise. Second, and more importantly, there are a good number of metrics which no longer exist in REST 10. You can find these in the Product Guide where footnotes indicate the changes as below.
So do these changes have any visible impact? Well, yes in fact they do. If you are using our current arrays with ESA, then upgrading to REST 10 (i.e. Unisphere 10) would deprive you of these metrics. Now, are there other metrics to take the place of these deprecated ones? Yes, mostly that is the case, but you would need to make the new association between them using the Product Guide. One area, however, where the change is apparent is in the PowerMax Overview dashboard.
In the PowerMax Overview dashboard there are a number of these REST 9 metrics used. Development decided to leave them and simply add the new metrics. Personally, I would have preferred they created new widgets for REST 10 and labeled the old ones so customers would understand why there was no data. But the way it is, when the old metrics are chosen, your widget will be empty as seen below when I select the metric Full (%):
Because of the way vROps operates, fortunately, this doesn’t cause any errors. The data simply isn’t there. So just be aware that there are metrics in the widgets which may show gray when chosen if using REST 10 – save yourself the SR. Some customers may choose to upgrade to Unisphere 10 even if they have the current 2000/8000 PowerMax platforms, but since the array software itself cannot be upgraded, embedded Unisphere will continue to run REST 9 which is still supported. Therefore, you have options. Remember you can always customize the dashboard to remove the metrics which no longer exist.
So the features.
The support of PowerMax File includes many new metrics along with changes to the dashboards.
One caveat to be aware of is that PowerMax File can only be managed with an embedded Unisphere. It is not available in an external Unisphere. Therefore if you have the new arrays, be sure to use your embedded Unisphere when adding an adapter.
For example, the PowerMax Overview dashboard has a number of new widgets:
Similarly in the PowerMax Metrics, some new File objects are present:
And PowerMax Topology:
In addition to dashboards, there are new alerts and views/reports for File.
Since NVMe/TCP is about how the device is presented, instead of a completely different datastore like NFS, the same metrics for FC/iSCSI apply; however, code changes were still required to make it work. The image below is a device presented via TCP, but other than the naming you can see there is nothing that distinguishes it from other datastores or devices. There are also no additional objects for NVMeoF.
As the upgrade process for ESA is is quite simple these days, it should be no issue moving to 7.0 if you wish. Since we support both REST 9 and REST 10 you don’t have to manage the adapter instances post install unless you have the new PowerMax 2500/8500.