Sorry I’m a little behind on this post – at EMC World in Vegas so besides the activities surrounding that, I’m 3 hours removed from the East Coast. But speaking of EMC World if you read my post and are out here at EMCW and want to shoot the breeze you can catch me at the VVol session on Tues or Thurs or at the EMC booth tonight from 7:15-9:15 PM at our opening.
Back to the product at hand. So ESA 4.0 went GA this morning with the following new/changed features:
-
-
- Supports EMC Avamar®. (see Steve Hegarty for info on that platform with vROps Avamar – Hegarty)
- Supports EMC Unity™ and EMC UnityVSA™ storage systems.
- Supports VMAX® data collection through Unisphere® for VMAX REST API instead of SMI-S Provider.
- Changes have been made to the VMAX Overview, Metrics, and Topology dashboards. The Top-N VMAX dashboard has been removed.
- Supports upgrades from EMC Storage Analytics 3.5.
- Some XtremIO alerts are disabled by default to align with XMS settings.
- Supports service level alerts for VMAX All Flash and VMAX3 arrays.
- VMAX2 (VMAX 10K, 20K, 40K, 10K File) arrays are not supported in this release.
- The vVNX VVOL Technical Preview is no longer supported as of this release.
-
For VMAX our big change comes in how we collect metrics. Prior to this release we gathered metrics through the SMI-S Provider. Starting with 4.0 we now utilize the Unisphere for VMAX REST API. The REST API is part of Unisphere for VMAX (U4V) so that is the software you will need to install before using ESA 4.0. Most VMAX customers use U4V so this isn’t an issue. Unfortunately you cannot install the REST API as a standalone product. Once U4V is installed, you will need to register the array to collect performance metrics. This is a straightforward process. Login to U4V and follow the steps in the image below. Your system administrator may need to do it for you (assuming it is not done already).
Click to enlarge – use browser back button to return to post
Once complete you can configure the VMAX adapter instance. Note that despite the change to use the REST API, we did not change the GUI interface since the field where the REST API is assigned is labeled generically as “Management IP”. For ESA 4.0 the field should contain the IP of the U4V instance along with the port. Be sure to change the credential to match U4V – the user needs to have a minimum role of PERM_MONITOR to collect metrics, but STORAGE_ADMIN to enable user actions (e.g. changing the SLO for a storage group). Here I am using the default user smc/smc. Always test the parameters once configured.
Click to enlarge – use browser back button to return to post
Alternatively if you are not configuring a new adapter instance but rather are upgrading from an earlier release, you will need to change the existing Management IP to U4V. After upgrading, your adapter instance will be unable to collect metrics and the adapter instance will show “Object down” like in the image below:
Click to enlarge – use browser back button to return to post
Simply edit the VMAX adapter instance in question and change the parameters. Note you should NOT remove the existing VMAX adapter instance. It is unnecessary; rather simply change the IP from SMI-S to the U4V address and port, test the parameters, and save.
The change to the REST API does have some implications for the type of metrics we collect. Foremost among these is that we no longer report device metrics. This is in line with our model of keeping everything at the storage group level. If device metrics are essential to your business, you will need to remain at ESA 3.5 and the SMI-S Provider. Note too that because of the removal of device metrics, ESA 4.0 no longer has a Top-N dashboard.
There are 2 other reasons you may need to stay at ESA 3.5. The first is if you run a VMAX2. ESA 4.0 marks the end of support for VMAX2 arrays. The second reason is if your VMAX3/VMAX All Flash does not have RDF directors (i.e. if you have not purchased SRDF), you will need to remain at ESA 3.5 because of a bug in the REST API. ESA plans to have a hot fix available in the near future to avoid the bug.
**** The SRDF issue was resolved in ESA 4.1 ****
Finally there are some minor changes to the VMAX dashboards – adding in the adapter instance resource for the metrics and topology dashboards and adding in SRP widgets to the VMAX Overview dashboard to replace the thin pool widget now that VMAX2 is no longer supported.
The documentation and software can be found here:
What’s the RDF bug in the REST API?
An unexpected error code is returned so ESA cannot handle it and thus it requires a code change which will be in the next ESA release.