Re: SSD/HDD Performance & Balancing

Sue,

Yes, since I don't have remote SST access, they only provide screen shot
"images", so I re-keyed data into excel (should have removed formatting).
The data comes from display device parity status.

574E DC06 is not referenced, & 574E DC07 shows twice, driving two raid sets
of 6 drives, and 12 drives. Why 6 & 12 ?

If you/we could start over, how would you recommend these 18 drives (same
model/size) be paired, grouped or parity'd ?

We will trace at twice a day, peak times, then balance low priority at an
opportune time, and ensure it's completion.




On Wed, Feb 25, 2015 at 5:11 PM, Sue Baker <sue.baker@xxxxxxxxxx> wrote:

Joe, something odd happened with the formatting and I can't
quite see what you posted clearly. However, it almost looks
like for the 574E, you are not using both DC06 and DC07 as I
only see a single controller resource name (DC07) in the list
you included with the parity set IDs and drives. I also see 2
parity sets with different number of drives for that controller.
This may be contributing negatively to your performance. But it
does not explain how things get worse the closer you get to
Friday each week.

Regardless, if you decide not to go with moving specific things
to SSD, I would encourage you to trace more times per week
followed by the balance. We have seen many customers have
success with multiple traces when they were struggling with a
single trace.

Another alternative depending on how busy your system is
overnight .... Maybe tracing for an hour each day (or maybe
twice a day, once in the morning and once in the afternoon)
followed by a balance each night may produce better results for
you. I'm thinking

STRASPBAL TYPE(*HSM) ASP(1) TIMLMT(180) PRIORITY(*LOW)

to make sure moving the data is less likely to get in the way of
getting real work completed.

Sue

Joe Wood <wood3875@xxxxxxxxx> wrote on Wed, 25 Feb 2015 23:20:03
GMT:

Sue,

Answers in reverse order and thank you.

Workload snapshot in my opinion is too short and creates some
of our own issues.

I'm saying we're not mixed.

ASP 1 SSD's have their own controller.
ASP 1 HDD's have their own controller.
ASP 2 HDD's (CEC), Dev partition 2, own controller.


Controllers:
Resource Name Type-Model Status
DC01 57CD-001 Operational
DC06 574E-001 Operational
DC07 574E-001 Operational
DC08 2BD9-001 Operational
DC09 2BE1-001 Operational



Parity Resource Hot Spare Set ASP Unit Type Model Name
Status
Protection 1 57CD 001 DC01 RAID 5/Availability N 1 1 58B2
109 DD002 Active 1 4 58B2 109 DD004 Active 1 3 58B2 109
DD003 Active 1 2 58B2 109 DD001 Active 2 2BD9 001 DC08
RAID 5 N 2 22 19A3 099 DMP051 Active 2 25 19A3 099 DMP010
Active 2 24 19A3 099 DMP024 Active 2 27 19A3 099 DMP050
Active 2 23 19A3 099 DMP053 Active 2 29 19A3 099 DMP047
Active 2 28 19A3 099 DMP045 Active 2 26 19A3 099 DMP055
Active 3 574E 001 DC07 RAID
5 N 1 17 198D 099 DMP030 Active 1 30 198D 099 DMP040 Active
1 20 198D 099 DMP035 Active 1 21 198D 099 DMP037 Active
1 18 198D 099 DMP031 Active 1 19 198D 099 DMP033 Active 4
574E 001 DC07 RAID 5 N 1 5 198D 099 DMP013 Active 1 16 198D
099 DMP018 Active 1 12 198D 099 DMP016 Active 1 7 198D 099
DMP003 Active 1 6 198D 099 DMP005 Active 1 10 198D 099
DMP025 Active 1 9 198D 099 DMP015 Active 1 11 198D 099
DMP007 Active 1 15 198D 099 DMP001 Active 1 13 198D 099
DMP012 Active 1 14 198D 099 DMP021 Active 1 8 198D 099
DMP028 Active

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.