Basic questions un-answered by storage scope?

Not to be rude but Control Center seems to be a low priority for EMC. We continue to hear that things will be fixed or updated in future versions yet every release seems to be more about new features and not fixing what is broken or only fixing Symmetrix related stuff. Control Center continues to be Symmetrix focused, Look at how many fields are not reported on for anything other than Sym. Things like any good level of Clariion or Celerra support seem to be of no priority.

The "future release" is the release that never comes. If a date can't be put on it then clearly it is not a priority. Just off the top of my head here are the things that ECC can't handle:

• The vmware agent can not support iSCSI to a Clariion
• No support for storage usage for VIO (AIX)
  
• No support for NPIV
• Celerra information is not integrated with backend storage
• No support for VMware via NFS to a Celerra
• Weak Clarrion management, no support for Clariion replication mgmt.
  
• No actual Celerra mgmt, Celerra replicator, etc
• Reporting is a mess as seen by this thread

So I had a chance last week to sit down with some folks in the Ionix group.  Had a full morning session with them including Gary Roberts, who is the Ionix Storage "Principal Product Manager".

These guys are running around the country to customer sites (two stops in Columbus Ohio last week).  Mostly based on peformance but they are putting serious effort into getting these tools squared away.

Any thing I brought up, including reporting for business decisions was very well recieved and they picked my brain until *I* was tired of it (some may have noticed I'm pretty passionate about it). heh.

My biggest take away is they understand that we need a reporting tool that is *not* by engineers for engineers.

It was well recieved.

My suggestion for everyone, KEEP THE PRESSURE UP.

If they don't know, they can't fix.

Be specific in your questions.

On another note, I'll be posting some queries that I wrote.  Will not work for everyone but maybe we can get some good stuff going here and someday, we'll get this thread closed as "answered".

bthessel - you're not being rude, we very much want to know what you think.  Thanks a lot of writing!

I can't argue with the fact that Symmetrix gets the most attention.  It has been the focus area for the last few years primarily because most (but not all) of our customers have them and expect all the latest features to be fully supported.  Things like VMax and FAST have required a lot of changes.

Support for non-VMware OS virtualization solutions is something we have been watching closely.  There aren't any standards and we are presented with a collection of OS-specific APIs that return wildly different things.  But honestly, what we're waiting for is customer adoption to get to the point where there is an obvious #2 player (behind VMware.)  As things stand right now, we're looking at either Hyper-V or Solaris Containers before VIO, but your feedback influences that.

Is there much iSCSI usage? I ask honestly, because I think we're going to shoot for FCOE and not iSCSI.

NPIV is something I'm personally concerned about.  How are you using NPIV today and where do you see it being used in the future?

I don't mean to ignore the other things on your list.  They are all areas we know we are short on.  We do have to prioritize based on the overall customer requirement.  The next big thing we are taking on is to rebuild the discovery engine so that it is agentless, requires fewer servers to deploy on, and handles the dramatic increase in scale our customers are experiencing.  We're also moving as much of our device discovery as we can to SMI-S and other standards so that we don't have to have unique code for every system.  Once we have that in place, increased support for something like Celerra becomes much simpler.

Scott

Edit: I removed my questions - I do want to know the answers but they don't belong on a support forum nor in this great thread about Storage Scope.  Let's move this discussion over to the ControlCenter Online User Community as Allen suggested - it's where we have detailed technical discussions on ControlCenter and the next-generation we are building now.

Starting with DMX's, at the most basic level.. I use this for "how much did I buy" number.

This will run against all frames, so you will need to filter by DMX (frame type, model or serial, whatever).  Also reports in Gig. 

Goto your query and hit the "manage calculated columns, create a new column.

Note: Our DMX's are pretty vanilla.  Two hyper sizes, fully configured.  No virtual as of yet.

Assuming usable space is configured space minus system devices.

Tier is always user defined, so you can create either a custom field or find out what is unique the the individual frames.

For instance, my "Tier 1" qeuery uses all arrays, but I filter by serial for 00019 (which is unique to my DMX-4's).

My "Tier II query" is identical to the above for Clariions, but change the filter to "APM".

We have user defined fields (UDF in the calculated columns) that denote location, again, we can filter for those as well.

Tier 1 query: (note this queury will not work in your environment.  Reasoning below).

SELECT DISTINCT srmarrayarray.arrayserialnumber arrayserialnumberarray, (decode(1048576, 0, to_number(null), (srmarraymetricsarray.arrayconfigtotal-srmarraymetricsarray.arraysystemresource)/1048576)) CustomDefinedColumn0, (decode(1048576, 0, to_number(null), srmarraymetricsarray.arrayalloctotal/1048576)) CustomDefinedColumn1, (decode(1048576, 0, to_number(null), srmarraymetricsarray.arrayunalloctotal/1048576)) CustomDefinedColumn2, (decode(1048576, 0, to_number(null), srmarraymetricsarray.arrayconfigtotal/1048576)) CustomDefinedColumn3, (decode(1048576, 0, to_number(null), srmarraymetricsarray.arrayunconfigtotal/1048576)) CustomDefinedColumn4 FROM srmarray srmarrayarray, (SELECT srmarraymetrics.* FROM srmarraymetrics, latestetlload WHERE srmarraymetrics.datekey = latestetlload.datekey) srmarraymetricsarray, srmudfinstance udfarraysarray WHERE (srmarrayarray.iscurrent = 'Y' OR srmarrayarray.iscurrent IS NULL) AND srmarrayarray.arraykey = srmarraymetricsarray.arraykey(+) AND srmarrayarray.arraykey = udfarraysarray.mokey(+) AND ( srmarrayarray.arraytype = 'Symmetrix' ) AND srmarrayarray.arrayserialnumber like '%00019%'

I have a seperate managed calculated row, so above where you see "CustonDefinedColumn0", that is really Configured Capacity minus Systems devices.  This is the not kidding actual usable space in the frame.

The reason for the "decode(1048576" is to get from Kilobytes from terabyes.

Regarding  " We also provide tiered storage by array or by site for capacity planning purposes. "

My ECC is at UB7 and I don't see tier "by array".  The Service Level only shows "Array Type", not Array Name.

Please advise how I can setup a tier by array.

Thanks.

We also provide tiered storage by array or by site for capacity planning purposes.

=>  How to do it?   I don't see tier by array.   I have 2 DMXs and the tier  "Symmetrix"  would include both of them.

This is actually near and dear to my heart as it's what spawned all the reporting issues.

About a year ago, we had like 9 tiers of storage.

(this is all opinion based, take it for what it's worth).

1: DMX standard r1bcv r2 r2bcv

2: DMX standard r1 r2 r2BCV

2a: DMX with 7200rpm drives

.

.

Etc etc.

Now we have

DMX=tier 1.

Clariion Tier 2.

NAS Tier 3.

Centerra Tier 4.

My personal opinion is that spindle speed is only half the equation.

A 300g 10k drive in a DMX is not technically faster than a 300g 10k drive in a clariion.

What is different is the cache size/caching algorithms, active active vs active passive, back end speeds, blah blah blah.

If a 10k drive is 40% idle.. if you make it a 15k drive, same size, now 60% idle...

Hi,

I am surprised by your quick reply and many thanks.

I define tier in "service level", is it the correct way?

If not, where do I define tier and how do I create a custom tier?

Thanks.

==================

1111 Marcus Avenue Suite M04

Lake Success, NY 11042

(O)516-734-3994 (Cell)516-232-3584

How do you define tiering?

Within the same array, just different spindle speeds etc or Sym=Tier1, Clariion=Tier2 etc...?

With all the new technology that is about to dawn upon us...with the new V-MAX arrays, FAST technology etc, etc you going to have tiering within the array from Flash down to SATA...also what about configurations...Mirror to Raid5, Raid6...all classified as tiering...and then we have virtual provisioning...

Is the reporting going to be able to keep up with all this...

Currently we have to report on many "things" to management regarding the storage - from RAW to configured, to raid type, to spindle speed (lacking in CLR's reporting ), to host allocations etc, etc...from Primary to SRDF to Clones (also lacking in CLR's reporting) etc, etc...

How do I run the customized SQL?  (in the Query, I can only view SQL but cannot modify it.

Thanks.

RichardYang wrote: How do I run the customized SQL?  (in the Query, I can only view SQL but cannot modify it. Thanks.

There are a couple of optoins depending on what you are trying to accomplish.  You can copy run a limited SQL select queries by creating a custom dashboard snapshot interface, though it's primary use case is to for customizing the StorageScope dashboard on the main screen.

You can also use many reporting tools, including Crystal Reports and Excel to connect directly to the StorageScope repository and run whatever SQL you like.  The schema is well documented and you can always start by looking at the SQL from a Query Builder query.

in StorageScope API and Repository reference guide,   Table "SRMHostByGroupMetrics", 

there is a field "HostArrayAccessibleRaw".

which Query can  display this field?

My problem is that we have hosts attaching to both DMX and CLARiion, and the HOST metrics seems to combine the RAW and there is no way to tell how much raw space used on each Array.

Column "HostArrayAccessibleRaw" in Table "SRMHostByGroupMetrics" can be displayed through custom query. In query builder, there is a build-in query called "Hosts in a Group". You can create a custom query by directly copying this query and then add column Hosts.Raw Accessible, which is HostArrayAccessibleRaw.

For your second question, If you have host accessing multiple arrays and you want to see capacities from each array, you can create custom query by following steps below:

1. Go to query build, click Create

2. Enter the query name you want and then Next

3. Select objects: Hosts, Host Arrays and Arrays, then Next

4. Choose column names, for example, host name, array name. For capacities from each array, please choose columns from Host Arrays. These columns will give you capacities from each array.

5. Add filter if you only want to see one or certain hosts.