[clug] Thanks eveyone - Storage and networking for Storage as used for supporting [Enterprise] virtualized environments

[George at Clug]

Scott, 

I realise you like people to check with google first, which I have. 
If I wanted to ask Google, I would not have posted on the CLUG mailing
list.

I was asking people who read the CLUG mailing list what was their
personal experience (if any) with "READ and WRITE performance figures
for real world production RAID configurations (RAID type &  number
of disks", should they be in a place to a) know, and b) able to share
this information.

Apologies if it was not clear as to who I was directing my question
to.

My personal experience maxes out around 8 x 7200 RPM Enterprise class
HD RAID 6 with live spare, so effectively data spread over five
drives, two parity, one live spare. I no longer have access to the
results from testing before putting the system into production, but as
I recall I was not impressed, had hoped for better performance.  I
do not recall the memory size of the RAID device's cache.


On Tuesday, 20-08-2019 at 10:27 Scott Ferguson via linux wrote:
> 
> 
> On 8/20/19 9:16 AM, George at Clug via linux wrote:
> > Thanks to everyone who was able to respond to my questions.
> > 
> > 
> > A remaining question, is anyone able to provide READ and WRITE
> > performance figures for real world RAID configurations (RAID type
&
> > number of disks) ?  
> 
> They'd be SSD disks right?
> 

The type of drive would depend on the type of drives that were used in
the person's RAID systems.

I would like to know RAID type, number of drives, and performance for
both SSD, 7200, 10000, and 15000 RPM Hard Disk drives (and/or any
other type of drive used in production RAID systems).

I do not RAID consumer grade drives.

>
https://www.google.com/search?q=READ+and+WRITE+performance+figures+for+real+world+RAID
> First result, also the Tom's Hardware links
> 
Sadly I have not as yet found any information that I can gain a good
grasp of for real world configurations via Google.
https://en.wikipedia.org/wiki/IOPS
https://www.techrepublic.com/blog/the-enterprise-cloud/calculate-iops-in-a-storage-array/
https://blog.storagecraft.com/raid-performance/

FYI: the first URL that Google offered me was
https://pdfs.semanticscholar.org/e1a3/83b89fb6af93e09eaec9fec2068571f997e9.pdf


The first URL that Google offered me regards Tom's Hardware was
https://www.tomshardware.co.uk/ssd-raid-benchmark,review-32689-3.html
" ..the RAID 0-based setups, consisting of two 128 GB and two 256 GB
SSDs..."
"...Samsung’s specs, the 128 GB 840 Pro..."

A RAID of only two drives?  Not something I would expect to see in
too many production systems.

Personally I would not want to be using consumer grade SSDs for
commercial production RAID environments. 
https://www.samsung.com/au/support/model/MZ-7PD256BW/  
"...Samsung’s specs, the 128 GB 840 Pro..."
https://www.itprotoday.com/high-speed-storage/there-really-difference-between-enterprise-and-consumer-solid-state-drives
"there is still a considerable difference between the price of a
consumer-grade solid state drive and an enterprise-grade solid state
drive"
https://www.ifsecglobal.com/uncategorized/what-raid-configuration-should-i-use/
"many SATA drives are built on the mechanics of desktop drive
technology and will not perform well within a 24/7 environment"

https://www.diffen.com/difference/RAID_0_vs_RAID_1
RAID 0 offers no redundancy and instead uses striping, i.e., data is
split across all the drives.

>From previous reading, I have a belief that some Enterprise cloud
storage providers use at least three separate RAIDed servers to keep
data on, each RAIDed using multiple (24?) RAID0 configurations. Since
RAID0 has no redundancy, if one server has any drive fail, that server
has totally lost all data. The hope is that not more than one other
server in the the three will have lost a drive until the failed server
can be returned to service.

https://storageservers.wordpress.com/2015/11/10/difference-between-storage-iops-and-storage-throughput/
Throughput- Data transfer speed in megabytes per second is often
termed as throughput. Earlier, it was measured in Kilobytes. But now
the standard has become megabytes.
IOPS- The time taken for a storage system to perform an Input/Output
operation per second from start to finish constitutes IOPS.

I did find this below link, not that it really helps with a clear
understanding of 16 and 24 disk arrays. I am guessing that since data
is still written to disks in 8 bit bytes, then performance does not
improve for RAID 6  after 8 drives for data + two parity?
https://support.hpe.com/hpsc/doc/public/display?docId=emr_na-a00017538en_us
End-to-End performance figures using virtual storage
MSA 2050 RAID 6 Hard Disk Drive (HDD) random results: Dual Controller
configuration, (180) 15K HDD, 10 drives per disk group, 9 disk groups
per pool, 9 volumes per pool.



> 
> > 
> > 
> > George.
> > 
> 
> Kind regards
> 
> -- 
>     A: Because we read from top to bottom, left to right.
>     Q: Why should I start my reply below the quoted text?
> 
>     A: Because it messes up the order in which people normally
read text.
>     Q: Why is top-posting such a bad thing?
> 
>     A: The lost context.
>     Q: What makes top-posted replies harder to read than
bottom-posted?
> 
>     A: Yes.
>     Q: Should I trim down the quoted part of an email to which
I'm reply
> 
> http://www.idallen.com/topposting.html
> 
> -- 
> linux mailing list
> linux at lists.samba.org
> https://lists.samba.org/mailman/listinfo/linux
>