EVA5000 Implementation Whitepaper?

[Abdul Rahiman]

Hi,

We are looking to consolidate about 4TB worth of storage (currently on various SCSI based Storageworks - HSZ70s, HSG80s etc. DASD arrays) into a SAN involving about 5 TRU64 servers. Can anyone point me to a good info on sthe subject in the form of a whitepaper/case study etc.. or a good EVA documentation to begin with ?

thanks,
Abdul Rahiman.

[Hemanth Gurunath Basrur]

Hello Abdul,

You will find all the documentation on EVA5000 from

http://h20000.www2.hp.com/bizsupport/TechSupport/DocumentIndex.jsp?contentType=SupportManual&docIndexId=3124&locale=en_US&prodTypeId=12169&prodSeriesId=321347

Also, go through these Websites:

http://h18004.www1.hp.com/products/storageworks/enterprise/index.html

HTH.

Regards,
Hemanth

[H.Merijn Brand (procura]

Hi Abdul, I'm just hijacking this thread to answer your question in Mark Syder's congrat thread, so you can decide yourself what to do with it :)

As of about a month ago, all stats are explained on my ITRC site. A reference to the explanation is also provided in the first post of the monthly statistics I post in the general category (just follow my link by clicking on "procura", and then "his questions")

My HP ITRC site pages can be found at (please use LA as primary choice):

USA Los Angeles http://mirrors.develooper.com/hpux/
SGP Singapore https://www.beepz.com/personal/merijn/
USA Chicago http://ww.hpux.ws/merijn/
NL Hoofddorp http://www.cmve.net/~merijn/

From there go to "Statistics" (upper right corner for the categories), and then on the "here" link in the first paragraph that should bring you to http://mirrors.develooper.com/hpux/stats.html

Enjoy, have FUN! H.Merijn

[Daniel Chow]

I don't know where the exact white paper is, but there was a case study done by Ken Bates on how to properly configure a EVA for certain applications, it provided good benchmark numbers and expected throughput. If you want I can fax you a copy, send me your fax number to my e-mail (dchow@becu.org) If your TRU64 servers are going to be Oracle DB servers then he recommends a 96 spindle disk group, anything more is a waste. General best practices:
- Have at least two logical disk groups, with double threshold set for each disk group.
- Dependent upon the size of your EVA you can spread the I/O by creating virtual disks from 2+ diskgroups (that way you can use all the spindles in your EVA) and then do simple addvols to your ADVFS domains
-If you have to "move" you data through addvols and rmvols, monitor your ISLs-if you are going through a fibre switches since they will be your bottleneck. I had to move 6TB from EMA to EVA and it took about a day or two but required no downtime for my Oracle DB--should still do it during less loads on your arrays.

Good Luck and congrats on your new EVA.

[Hemanth Gurunath Basrur]

Hello Abdul,

Refer to the foll. case study on how Pekin Hospital used EVA 5000 from the link below:

http://h30046.www3.hp.com/casestudy.php?pagesite=HEALTHCARE&topiccode=20040330_PEKIN_HOSPITAL

HTH.

Regards,
Hemanth