HP MSA 2060 with VMWare 8 - Space Question

[kmccb]

Hello, I have a simple small 2 node ESXi 8 cluster consisting of:

(2) HP ProLiant DL380 Gen10 Plus direct connected to an HP MSA 2060 using iSCSI. The MSA has (12) 1.8 TB SAS drives.

I have created 1 pool and disk group utilizing all 12 disks in HP's MSA-DP+.

I then created (3) Volumes and presented them to my ESXi hosts.. The 3 volumes are:

SAN Volume 1 - 1,999.9 GB <--> VMWare Datastore 1 - 1,820 GB (180 GB Difference)

SAN Volume 2 - 3,299.9 GB <--> VMWare Datastore 2 - 3,000 GB (299 GB Difference)

SAN Volume 3 - 7,999 GB <--> VMWare Datastore 3 - 7,280 GB (719 GB Difference)

All 3 datastores in VMWare are VMFS 6.82 / 1 MB Block Size.

I have a few questions..

1: Why am I losing 1.2 TB from the MSA to VMWare?

2: The current volume usage per the MSA is: 9.3 TB Allocated // 3.9 TB Total Committed // 1.0 TB Unused. What exactly are these numbers comprised of?

3: How much free space on the SAN can I still present safely to VMWare?

Thanks in advance!

[JonPaul]

@kmccb 
This is most likely due to the difference in GB vs GiB.
All drive manufacturers present their products in GB (Base 10) 
Most, if not all, OS manufacturers present storage in GiB (Base 2)
Back when this started HDD sizes were in the MB range and difference was hardly noticable.  Now that HDDs can be 10s of TB there is a big discrepancy.
The MSA has to bridge the 2 and by default lands on Base 10.  If you change your console settings or WebUI setting to display in Base 2 you will likely see the same capacities which the OS displays (minus some small capacity for MSA overhead)

Allocated: How much capacity is allocated to volumes   Your pool has ~10.3 usable TB,  9.3 is currently allocated to volumes. If every volume reached maximum capacity your system would have 9.3TB of data.
Total Committed:  How much actual data is in the Pool.  Of your 9.3TB of volume capacity, only 3.9TB are used. The rest of the capacity is waiting for data.
Unused:  How much physical capacity remains if all volumes are full.  This is a guage to how close you are to an overcommitted/over-provisioned/thin-provisioned system

Define safe...  Your system has 1TB of capacity which can be allocated to a volume(s) before you are overcommitted.  If you are using snapshots, then you could run into space problems of creating a snapshot due to lack of capacity even if you are not overcommitted.  If you are not using snapshots then allocating up to the Pool capacity will allow you to use all of the available pool capacity.  You may want to look at the Pool setting for 'over-commit' to avoid allowing volumes capacity more than actual capacity.   If you do overcommit the Pool, more volume capacity than physical capacity, then the end user could hit an unexpected 'out of storage' message where the OS will tell them that they have capacity but will error that they are out of capacity when something is attempted to be written.  If you are overcommitted the system will warn you as the system runs out of capacity.

[kmccb]

Thank you for the reply.. Gotcha on the GB vs GiB, appreciate the explanation there..

Still have some followup regarding my other questions..

The "Total Size" for my pool is 14.3 TB.. I created volumes that total 13.3 TB (2 TB+ 3.3 TB + 8 TB).. You said allocated is 9.3 TB so I'm confused on that..

You also mentioned that total committed was how much data is in the pool.. 4TB is what the MSA says is committed.. I have a file server that is using 6TB of data alone so I'm also confused on that number.. 

Then back to my 9.2 TB allocated + 4 TB committed + 1 TB unused = the 14.2 TB total size, no?

Thanks for the help.. 

[JonPaul]

@kmccb 
A. I am wrong
B.  
Allocated - actual used capacity -- this should then make the File server using 6TB + other volumes make sense
Committed - should be the volume total capacity -- this is not aligning with what you are reporting and should be showing 13.3TB
Unused - Capacity which is not committed (used by volume capacity)

If you use the CLI it may be easier to see the allocated size per volume:
CLI>  show volumes
And available (not currently used) capacity in your pool:
CLI> show pools

The Web interface is using this information to combine and report out the usage seen.

[kmccb]

Thanks.. One more question.. Is there a calculator anywhere to show how big the pool would get by adding additional drive(s)?

I know there are the normal RAID calculators but I'm using HP's MSA-DP+ RAID..

Thanks

[JonPaul]

@kmccb 
Best calculator is the MSA Sizer  (Ninja Online)  https://ninjaonline.ext.hpe.com/
Use the non-recommended manual configuration and plug in numbers of drives.
In rough numbers the capacity will be:   (Total Capacity of All drives - capacity dedicated to spare) * (1 - % reserved for parity) == usable capacity
With MSA-DP+ the parity is 20% (0.20)  (RAID 6,  8D + 2P)

[kmccb]

OK thanks. That Ninja Calc helps a lot.. The model drives in the SAN are ST1800MM0129.. Can I just order that model from anywhere and pop them in the SAN or do they have to be ordered directly from HP?

[JonPaul]

@kmccb 
It would be recommended to only use drives supplied by HPE.
1. The drive carrier is not a saleable item from HPE so foreign drives would not fit
2. Drives from HPE are configured to work in the MSA,  format or other parameters could render the drive UNUSABLE in the MSA
3. Foreign drives may have a different version of firmware which will not be tracked and updates will not be available from HPE, Critical firmware issues can not be addressed
4. Warranty of the drives would be problematic if one should fault

[kmccb]

Thanks, I figured as much.. Is there a link to buy them online or?