SSD Storage for E1

cncjay

Well Known Member
Hello All,

We are looking to replace our existing storage system being for JDE.

Just wondering, has anyone implemented an SSD based solution within the last 6 months or so that they can recommend?

Our production database is around 2TB on SQL 2008R2 platform

Not being a network/infrastructure guy, not sure if that question is too vague but if there are some storage best suited for JDE, I'm interested in knowing that.

Regards,
CNCJay
 
While not directly comparable, we use some SSD on our IBM i installation. I have placed our most active tables and their indexes there. It has a dramatic effect on performance for us.

Tom
 
Hi Jay,

We are using SSD on our IBM i with JDE 9.1. We have most system, server map, data dictionary, central objects on the SSD drives. We also put some high use control tables (UDC, menus) and business data (address book, G/L master, AAIs, item master, etc).
We did see significantly faster database access times with a corresponding increase in application response times and reduction in UBE processing times.
 
Thank you so much for your feedback. This will be our first venture into an SSD storage solution, looking forward to the changeover. I am hoping we will also see a significant performance gain.
 
SSD on System I

While not directly comparable, we use some SSD on our IBM i installation. I have placed our most active tables and their indexes there. It has a dramatic effect on performance for us.

Tom

I am curious if you saw any kind of performance improvement with the System I SSD. We went thru this as well with a VIOS implementation. We monitored the performance both with a tool and colloquially. Neither suggested any kind of performance improvement.
 
Yes, I was able to compare as we did a test failover. Our DR machine is identical to our PROD machine with the exception of the SSD. Performance was significantly slower (0-50%) depending on the process and what I had assigned to the SSD. We are on 6.1 so I have to assign whole tables/indexes. Processes for tables that were not on SSD ran about the same (0%) and processes with lots/all their tables on SSD ran up to 50% slower when on the DR machine.

Since the hardware otherwise was identical, I had both machines memory pools tuned the same. Mirroring ensured the rest was identical.

Tom
 
Back
Top