Tuesday, April 4, 2017

COLLABORATE17: Day 2

E-Business Suite on Oracle Cloud: An Architect's Perspective - Roy Hunter

Business priorities - innovation, agility, lower cost and risk (grow and transition the business)
IT priorities - on-demand self-service, rapid elasticity, measured service (enable business growth and transformation)
Five QA systems to support a PROD system, TEST, UAT, SIT, performance test, and stress test!
Migration tools allows automated cloning.
Reimplement onto SaaS
Re-platform onto IaaS - why?  Current environment tailored (customized) to support unique business processes and requirements, while SaaS requires business process re-engineering and SaaS modules lack features/functionality, or don't exist.  Doing IaaS allows you to eventually move to SaaS when ready while getting benefit of IaaS immediately!
Constraints on moving - Data Sovereignty, Control, Latency
Oracle Cloud Machine moves into your Data Center to solve some of those constraints.
Co-existence between Cloud and on-prem is possible, and actually even expected!
Can't get a trial account for Exadata, but if you talk to your tech rep they may be able to help set up a POC
https://blogs.oracle.com/EBSandOracleCloud/
This is for EBS 12.1.3 and up.
There is also a new disconnected OCM which to me sounds like an Exadata box, where it gets installed in your DC and then does not connect to Oracle.

E-Business Suite 11i/12.1/12.2 to 12.2.6 Upgrades: Quick, Fast, and Easy - @OATCInc

Relational calculus is what Oracle is built on, did you know?
Review of the support timelines, which is always good.
742060.1 DB support note
If you are going to do a fresh install, do R12.2 now, there is no reason to do R12.1 anymore.
1439822.1 EBS customers get updates for JRE
Forms won't launch in Chrome without an addon that is unsupported anyways!
Microsoft Edge doesn't support Java, so won't support EBS!
Oracle Forms are not going away due to how powerful they are, and cannot replicate what they do in Java...yet at least!
There are 22 mobile apps right now - 1641772.1
2114173.1 R12.2 Knowledge Center
1984710.1 architecture
UI login in 12.2.6 and favorites have a tablet look and feel which is completely different!
E48834-13 11i to R12.2 upgrade document
R12.2 patches are C patches, R12.0 are A, and R12.1 are B, so know which are which and do not apply the wrong thing!
How long it takes depends on: Time, scope, budget (famous triangle)
1 month per module!

Navigating My Oracle Support with the Customer Support Council - Doug Hahn - @doughahn

oaug.org/resources/csc
Enhancement Requests
Extended Premier Support for 12.1
ORAchk, Oracle Analyzers, Support Identifier Groups (virtual CSIs)
Hot Topic notifications
Bundle Analyzers - concurrent reports DO update
1939637.1
Focusing on being proactive

Architect Data for Excellence - Ian Abramson - @iabramson

Bi-model BI, this is new to me, so need to research it a bit.
Wal-Mart is like a cult…Sam Walton was focused on innovation
Hans Peter Luhn, Bill Inmon, Ralph Kimball - major influencers on Business Intelligence/Big Data
Ian deals with a 40 Petabyte data lake!
Variety, volume, velocity, veracity
Data Lake - highly agile, data scientists vs. Data Warehouse - less agile, business users
Need to celebrate failures as much as successes - this is a major takeaway because the thought process behind it is that you learn from failing just as much (if not more) than succeeding!
Key question is: What do we expect from Big Data?
DW is highly governed data, low latency, 1000s of users vs. Big Data raw and unstructured data, unrestricted and ungovernable data
Out of a million products, probably 200k are right, so how does that throw off the company?  Millions of dollars out the door!
What do we need to do today?  Need to provide access to all data, be able to bridge data between business units and data sources, keep data secure, bi-modal BI is the new analytic approach.
Hadoop is not a database.  Hadoop is not an IDW.
Data warehouses are not going away, they are going to instead evolve!
Big Data is not a solution, but a tool
New analytic needs - new data integration, business need for lower data latency, support for dynamic structures, report personalization.
How to evolve?  Understand the architecture.
Analytic Sandbox -> Raw Data Hub -> Enterprise Data Lake
Three components of data architecture: landing & assembly, general DW, specialized data warehouse

Low-Hanging Fruit: Tuning & Securing WebLogic Server - Sid Patel & Smarajit Pramanick - @ASTCorporation

Intro of WLS - example of a sandwich to show how what is in the middle might really matter…great example!
Common issues - system errors (stuck thread, thread waits, timeout, memory issues - StuckThreadMaxTime default is 600 seconds), bad throughput (server not responding to new requests), loss of messages (intractability of messages, network/socket issues), poor performance (request taking longer time than expected, SLAs violated)
Common performance issues - memory leak, thread deadlock, WLS session replication, high number of open file descriptors or sockets, long running DB queries, JVMs spending took much time in GC, too chatty system, improper hardware sizing, network bandwidth, payload size, load balancer issues, bad application design
Performance tuning - how is WLS performance measured?  Response time, latency, throughput, resource utilization - how is WLS performance tested?  Functional, stress, load, and soak testing
Benchmark
Identify possible bottlenecks
OS tools - netstat, iostat, top, vmstat
WLS tools - server logs, WLS diagnostic framework, WLS console & EM, WLST
JDK tools - JPS, thread dumps, heap dumps, Java flight recorder, jvisual vm, jconsole, jstat (some of these I do not know so I need to look them up)
Log and debug options - logging service, log subsystem events (startup/shutdown, application deployment, subsystem failures), communicate status, domain message log (overall status (server logs and out, domain logs)), change log location, message catalog (Java logging API), use Java logging (Log4j), subsystem logs (http, transaction, data source, diagnostic images)
Log severity levels
Log rotation policy
WLDF (WebLogic Diagnostic Framework) free out of the box
OS parameters need to be tuned, especially for Linux
file descriptor since everything is a file - "Too many open files"
ps -eaf | grep Java -> pfiles processID | grep rlimit
ulimit -H -n
ulimit -S -n
TCP tuning, default wait time netstat -a | grep TIME_WAIT | wc -l

1 comment:

  1. If you don't remember this, your car may be stolen!

    Imagine that your car was taken! When you approach the police, they inquire about a specific "VIN search"

    A VIN decoder: What is it?

    Similar to a passport, the "VIN decoder" allows you to find out when the car was born and who its "parent"( manufacturing plant) is. You can also find out:

    1.Type of engine

    2.Model of a vehicle

    3.The DMV's limitations

    4.The number of drivers in this vehicle

    You will be able to locate the car, and keeping in mind the code ensures your safety. The code can be checked in the database online. The VIN is situated on various parts of the car to make it harder for thieves to steal, such as the first person seated on the floor, the frame (often in trucks and SUVs), the spar, and other areas.

    What happens if the VIN is intentionally harmed?

    There are numerous circumstances that can result in VIN damage, but failing to have one will have unpleasant repercussions because it is illegal to intentionally harm a VIN in order to avoid going to jail or being arrested by the police. You could receive a fine of up to 80,000 rubles and spend two years in jail. You might be held up on the road by a teacher.

    Conclusion.

    The VIN decoder may help to save your car from theft. But where can you check the car reality? This is why we exist– VIN decoders!

    ReplyDelete