Exadata Cell Capacity and performance
Date: Tue, 17 Aug 2021 01:17:44 +0530
Message-ID: <CAKna9VYfk7qB7yT8r3rcBmv=YHmOwgsS9qLqk6peD-1vSppPAg_at_mail.gmail.com>
Hello All, We have one critical database(say DB1) moving from an X3-2 HC full rack exadata machine to a new X7-2 HC half rack exadata machine. The existing X3 full rack was hosting ~4 different other small databases along with that critical database- DB1. The new X7 will be hosting only the DB1 database for the time being.
So in this scenario as I see the data sheet and put the figures in as below it seems the X7 half rack is better with respect to both flash and disk storage capacity and flash IOPS. Just that the disk IOPS seems to be lower in case of X7 half rack i.e. ~28K vs ~18K. So considering we are not changing any application functionality of oracle software version etc, so is it okay to assume we will be seeing improvement only , thus its safe to directly go with the migration without any full fledged performance test? Or there is any chances of performance degradation here and thus a full fledged testing needed like real application testing?
IOPS For Half Rack
Flash Read IOPS Flash write IOPS Disk IOPS
X3-2 High Capacity (Full Rack) ~1.5million(~k per cell) ~1million(K per
cell) ~28K
X7-2 High Capacity (Half rack) ~2.3Million ~2.1million ~18K
Storage capacity For Half Rack
Raw disk Storage High Redundancy Normal redundancy Flash Storage
X3-2 High Capacity (Full Rack) ~504TB (TB/cell) ~TB(TB/cell) ~224TB(
TB/cell) ~22.4TB(~TB/Cell)
X7-2 High Capacity (Half rack) ~840TB (TB/cell) ~TB ~341TB ~179.2TB (25.6TB/Cell)
-- http://www.freelists.org/webpage/oracle-lReceived on Mon Aug 16 2021 - 21:47:44 CEST