Most Oracle DBA are sufficiently educated about benefits using large memory pages for Oracle database SGA to reduce overhead and improve performance. If you want to read more about it you can start from that Oracle blog or read it from other multiple articles and blogs. Oracle is using parameter use_large_pages to direct behaviour of an Oracle instance during startup.
In the previous versions before 19c we had three possible values – “TRUE”, “FALSE and “ONLY”. Since Oracle 18.104.22.168 the “TRUE” meant that the instance will allocate as many hugepages as free available in the system and get the rest from the normal small pages. The “FALSE” would tell it to not use the hugepages at all and the “ONLY” would be able to start an instance only if sufficient number of free hugepages is available in the system to fit all SGA in it. The “TRUE” was default for all databases.
In the 19c version we got one more value – “AUTO_ONLY” and now it is the default value for Exadata systems running Oracle Database 19c. The description in documentation is not totally clear and sounds very similar to the description of “ONLY” value. Here is an excerpt from the documentation:
“It specifies that, during startup, the instance will calculate and request the number of large pages it requires. If the operating system can fulfill this request, then the instance will start successfully. If the operating system cannot fulfill this request, then the instance will fail to start.”
Let me show you how it works. Here is my sandbox with a 19c database and no hugepages is configured on the box by default.
In the previous posts I shared my first impression and how to start using the Google Bare Metal Service (BMS). In this post I will try to show some numbers related to the performance of the solution and you can compare it with your existing environment.
Let me start from the box characteristics. For my tests I was using a “o2-standard-32-metal” box located in the us-west2 zone (Los Angeles) . The solution was configured with 2Gbps interconnect and had a couple of storage resources attached to it. The first one was represented by two 512Gb disks based on HDD storage where I placed my binaries and a recovery ASM disk group and the second was a 2Tb volume “all flash” I used for data. Here is summary table:
BMS Box type
Intel(R) Xeon(R) Gold 6234 CPU @ 3.30GHz
512 Gb – Standard disk
512 Gb – Standard disk
2048 Gb – All flash
4 NICs Speed: 25000Mb/s
Oracle Linux 7.9
BMS box characteristics.
Before starting the tests I updated my Oracle Linux and installed a number of packages required for my Oracle database and packages to test IO and Network such as fio and iperf3. Here is a summary table with software and tools used to test the performance.
Since the first days of working in the Google public cloud there have been debates about the possibility to move an Oracle workload to GCP. The major concerns were coming not from the technical challenges but rather from Oracle’s licensing policies and guidelines. In the famous Oracle’s document about licensing Oracle software in the public cloud it was stated – “This policy applies to cloud computing environments from the following vendors: Amazon Web Services – Amazon Elastic Compute Cloud (EC2), Amazon Relational Database Service (RDS) and Microsoft Azure Platform (collectively, the ‘Authorized Cloud Environments’)”. So the Google Cloud was not listed as an ‘Authorized Cloud Environment’ and it was unclear how to apply the Oracle licensing there. I believe it will be sorted in time but in the meanwhile as a solution Google presented a Bare Metal Service as the platform for Oracle workload.
Oracle Database as a service (DBCS) on Oracle Cloud Infrastructure (OCI) had been traditionally built based on Oracle Grid Infrastructure with ASM as the main storage layer for the database, however Oracle recently has started to offer a Linux LVM as the storage alternative. Which option is the better alternative? Let’s review some of the differences between the two options.