According to Exadata consolidation guide, this is what you can consolidate on Oracle specialized Hardware:
NOTE: The maximum number of database instances per cluster is 512 for Oracle 11g Release 1 and higher. An upper limit of 128 database instances per X2-2 or X3-2 database node and 256 database instances per X2-8 or X3-8 database node is recommended. The actual number of database instances per database node or cluster depends on application workload and their corresponding system resource consumption.
But how many instances are actually beeing consolidated by DBAs from all around the world?
I’ve asked it to the Twitter community
I’ve sent this tweet a couple of weeks ago and I would like to consolidate some replies into a single blog post.
who has done more than this on a single server? $ ps -eaf | grep ora_pmon | wc -l 77 #oracle #consolidation
β Ludovico Caldara (@ludovicocaldara) October 25, 2013
My customer environment however, was NOT a production one. On the production they have 45.
Some replies…
@CacheFlush @kevinclosson @ludovicocaldara I know one customer with 58
β Bjoern Rost (@brost) October 25, 2013
@ludovicocaldara I have more π β Wissem El Khlifi (@orawiss) October 25, 2013
@ludovicocaldara ay cant believe it, think they removed some DBs, i have now 73. Ok u win π β Wissem El Khlifi (@orawiss) October 25, 2013
Wissem cores 73 on a production system, 1TB memory!
@ludovicocaldara about 150 TB of storage, 1TB memory, 40 Cores …
β Wissem El Khlifi (@orawiss) October 25, 2013
@ludovicocaldara we have at least 4 servers like this , big ones! 2 for prods …
β Wissem El Khlifi (@orawiss) October 25, 2013
Chris correctly suggests to give a try to the new 12c consolidation features:
.@ludovicocaldara a good playground for either threaded_execution=true or multitenant π
β Christian Antognini (@ChrisAntognini) October 25, 2013
Kevin, as a great expert, already experimented one hundred instances environment:
.@ludovicocaldara up to 100 instances/host when researching for this paper: http://t.co/sIAuvnDJ2D Proof point stayed much lower though
β Kevin (@kevinclosson) October 25, 2013
But Bertrand impresses with his numbers!
@ludovicocaldara Will give you the right number on Monday but I would say around 120 π
β Bertrand Drouvot (@BertrandDrouvot) October 25, 2013
@ludovicocaldara Here is the exact number dude: ps -ef | grep -c ora_smon 131
β Bertrand Drouvot (@BertrandDrouvot) October 28, 2013
@BertrandDrouvot @ludovicocaldara damn your the winner, i canβt higher then 118 but this customer is still busy migrating more.
β Klaas-Jan Jongsma (@futureveterans) October 28, 2013
@ludovicocaldara capacity planning is a big challenge here with currently around 1800 databases and this is constantly increasing…
β Bertrand Drouvot (@BertrandDrouvot) October 25, 2013
@ludovicocaldara @orawiss 1TB is our default as of a few months π
β Bertrand Drouvot (@BertrandDrouvot) October 25, 2013
Intel platform with 1TB of RAM = Xeon E7, suggests Kevin:
. @BertrandDrouvot @ludovicocaldara @orawiss so you standardize on Xeon E7 ?
β Kevin (@kevinclosson) October 25, 2013
@kevinclosson how did you guess ? π @ludovicocaldara @orawiss
β Bertrand Drouvot (@BertrandDrouvot) October 25, 2013
. @BertrandDrouvot @ludovicocaldara @orawiss because there are no 1TB EP offerings… at least not until after Haswell. Ivy EX does 12TB π
β Kevin (@kevinclosson) October 25, 2013
Flashdba has seen 87 instances on a single host, but on a Multi-node RAC: but still huge and complex!
@ludovicocaldara @BertrandDrouvot @orawiss When I worked for Oracle I saw 87 instances on single a node – and it was six node RAC
β flashdba (@flashdba) October 25, 2013
@ludovicocaldara @BertrandDrouvot @orawiss It was a production system too – and yes, it was a nightmare!
β flashdba (@flashdba) October 25, 2013
@kevinclosson @ludovicocaldara @BertrandDrouvot @orawiss The cluster had around 120 DBs, not every database had an instance on every node
β flashdba (@flashdba) October 25, 2013
@kevinclosson @ludovicocaldara @BertrandDrouvot @orawiss The result was between 50 and 87 instances running per node #complicated
β flashdba (@flashdba) October 25, 2013
Conclusion
Does this thread of tweets reply to the question? Are you planning to consolidate your Oracle environment? If you have questions about how to plan your consolidation, don’t hesitate to get in touch! π
—
Ludo
Latest posts by Ludovico (see all)
- New views in Oracle Data Guard 23c - January 3, 2024
- New in Data Guard 21c and 23c: Automatic preparation of the primary - December 22, 2023
- Does FLASHBACK QUERY work across incarnations or after a Data Guard failover? - December 13, 2023
Yep, it’s not a single node RAC, rather a multi-node RAC where every database has only one instance started. The failover is automatic and the migration from one node to another is transparent to the applications. IMHO It’s ideal for small databases.
More info here: http://www.oracle.com/us/products/database/options/rac-one-node/overview/index.html
GI is the Grid infrastructure (the foundation for a RAC or RAC One architecture) and ASM is the “storage manager” from Oracle that replaces cluster filesystems delivering better performance and manageability.
Hope it helps
I’ve got the info about failover of single node RAC.
Thnaks for replying. My unerstanding was RAC is mostly used for high availability for critical applications. Our applications are mostly vendor based. With single node RAC, how we do achieve failover? βo-motionβ and GI+ASM are new to me. I will check doumentation and let you know if I have any questions.
Hi Ramki,
for sure! not only, I think it could be a FAR BETTER solution compared to VM/Linux, because you don’t have to take care of many and many operating systems but rather you can concentrate on your databases and make huge savings on storage and memory.
If you have to start from scratch with a new architecture, for many small databases I would suggest a RAC One Node cluster, so you can benefit from “o-motion” and leverage GI+ASM whilst reducing licensing costs of a full RAC solution.
If you’re on Oracle Standard, however, you can do a full RAC SE with one or more clusters composed by a maximum of two nodes each (4 socket total for each cluster). Different costs AND different limitations! π
Cheers
—
Ludovico
We are looking forward to consolidate 100+ small Oracle databases. Do you think RAC is still a viable option for consolidation compared VM/Linux?