Feature Request Thread

vSphere Path Selection Policy - please make it a selectable choice during datastore creation using the vCenter plug-in.  Or have a global setting somewhere that can be changed to set to Round Robin as the default.  Please!!!

Would be nice if this was in the vSphere GUI somewhere, but you can already do this: kb.vmware.com/.../search.do

I'd love to see Active Directory integration. It's quite odd that Compellent doesn't offer this feature yet (or maybe I just don't know about it).

Something that would be nice might be a revisit of the Enterprise Manager/Storage Center relationship. When we bought our systems we were pitched that EM was an EMS-type app. Well the application kind of fakes that functionality but really it just collects statistics and proxies some requests to the SCs. It doesn't make sense to me that the vCenter plugin is client-side only and requires you to add the IP addresses of the SCs directly. vCenter should be able to talk to the SCs directly, or maybe use something like a central EMS for us folks with multiple SCs...

I have the following suggestions:

1) It would be nice if Enterprise Manager evolved into a more vCenter like management interface. Specifically, the ability to create a standard set of replay and storage profiles that could be applied to the Storage Centers. It is no fun having to manually recreate those profiles as new Storage Centers are introduced.

2) Similar to request #1, the ability to replicate certain user configured settings to new Storage Centers would be nice as well. The ability to create a "Storage Center Profile" within Enterprise Manager that could then be applied to new systems. It would be useful in not only applying common Replay or Storage profiles, but also folder configuration, user and group configuration, etc.

3) Modifying some of the user capabilities within Storage Center, such as the ability to Copy/Paste (Volume folders, server folders, etc.) and the ability to drag/drop (moving volumes and/or servers).

4) AD Integration from both Storage Center and Enterprise Manger (as mentioned before). In addition, more granular control over Storage Center privileges. For example. Volume Managers do not have the ability to create servers and the only method to provide that ability is providing administrative access.

5) Along the same lines as request #4, it would be beneficial to be able to assign user rights to objects within Storage Center and not be limited by Volume or Server folders. We found the current method of assigning group access to folders conflicted with the Server/Volume folder strategy we originally wanted to implement because it did not jive with how we wanted user rights to be assigned.

6) Include the Replay and Mapping options when configuring New User Volume Defaults.

7) Within the rack templates, support for racks larger than 50 RMU would be nice for those of us with taller racks.

Yes, great ideas! These sort of things would make my life a lot easier and exactly along the lines of what I was alluding to.

We have experienced inaccuracies in the information reported by Enterprise Manager.   In particular we had situations where the remote ports don’t match up.  See case 116604 for more information.

I have a few that I would like to see in future releases.  The first has been stated a few times, but hopefully if enough of us ask it will happen.  AD Integration is big on my list.  We as others do have multiple controllers and multiple users managing them, and for compliance reasons it would be nice to know that when we create or disable accounts in our central directory, access is granted and removed automatically.

The second request is for some type of Quality of Service (QOS).  I am a huge fan of virtualized storage, and pooling resources across all the components of a system for the greatest efficiency of the system as a whole, but this can cause other issues.  On a busy system, it is very difficult to tamp down less important systems so that they do not steal IOPS and processing power from more important servers.  While there is the ability to limit systems to certain tiers, each time you go down this road you detract from the value of the virtual nature of the system.  For example, our typical configuration is to allow all LUNs to use all tiers, and then lesser LUNs to only use lower tiers.  However, since important systems also migrate data downward, systems at the bottom can still slow down systems at the top.  Conversely, if I then decide to keep my important systems only at the top tiers, I lose the value of the tiering and I've created the mess of a legacy SAN.

It would be useful to be able to set limits on certain systems or LUNs to ensure that they can not exceed certain performance thresholds.  

From a VMware perspective, this is possible with Storage I/O Control (SIOC), but this functionality is limited in what it can do, and is also on a VM by VM basis (lots of configuration points).  If we could keep all LUNs on all tiers this would maintain the efficiency and smarts of the automated tiering (the secret sauce of the Compellent) and having the ability to assign a "Class of Service" or "Quality of Service" on a LUN basis would help ensure that LUNs of lesser value could not cause a service degradation system wide.

The QOS recommendation is a really great idea. We've encountered the exact scenario described where certain volumes, based on their utilization, would spike to several thousand IOPS causing us to completely re-evaluate our hardware configuration. On one hand we were pleased with how well the system was able to perform after being migrated to the Compellent, along with the granular performance monitoring via Enterprise Manager. However, with this scenario occurring early on in our data migrations from our previous arrays, we were concerned with how IO starved other systems may have been and what they would be capable of after migration to the Compellent. We toyed with the idea of purchasing separate disks to create a "graveyard disk pool" (for lack of a better term) where we could place volumes capable of consuming significant disk resources but who's important or criticality didn't warrant it. As CGEHRING pointed out, this detracts from the concept of virtualized storage. The abiity to QOS volumes within Storage Profiles would be a huge benefit.

Seconded - we run into the same issue.  We only allow certain systems to use tier 0 (in our case SSD) but as data is quickly migrated down, it causes a lot of stress on tier 1 which cause all systems to slow.  I'd also like to see multiple daily data progressions available for SSD tiers.   

1.  AIX support with virtual port mode 

2.  Be able to change storage profile with right click or in edit volume 

3.  Charge back uses tier 1 price for comparison, that should be configurable as flash drives skew the numbers. 

4.  Cache information (% cache hits, etc) 

5.  Add search to EM 

6.  Replications shouldn't default to lowest tier 

I would like to give a +1 to JBrunt's following suggestions:

2.  Be able to change storage profile with right click or in edit volume  

4.  Cache information (% cache hits, etc)  - Blew my mind this was not there..  but then again CML is not a cache heavy device.

6.  Replications shouldn't default to lowest tier

The last one burned us...  if you think about reality, and how you often sell on the fact that your BC system does not need to be as rubust as your primary, you will crush the lowest tier of the BC site with all the changes happening at every tier of production.  It would be even cooler if the replication data could be tagged to land in the same tier as the source instead of just top down or bottom up.

Within the Charge Back module, it would be nice if the individual storage usage of VMs could be reported on.

In regards to JBrunt 1st request, I don't believe Compellent has any control over AIX not supporting Virtual Port Mode. That is due to AIX not supporting NPIV. I've heard AIX 7.1 may introduce the necessary support but I cannot confirm that.

Great thread, but it would be good to to get some feedback from a Dell/Compellent member so we know this is being watched and considered.  I see some common threads and good suggestions, so feedback on what is being considered or already in the roadmap would be good.

notify customers about upgrades and patches inside all storage center GUI apps when they are available.