 |
| |
| home >> speakers >> Larry Youngren |
| |
|
Larry Youngren
After more than 30 years of experience leading the design efforts for System i database and journal support at IBM, Larry Youngren recently retired from IBM and now lectures, writes, and consults on high availability issues.
For 30 years, Larry served as a microcode designer for the lower layers of the i5/OS operating system and frequently consulted with customers regarding High Availability and Journal performance issues. During his IBM career,he worked exclusively with the microcode, first for the S/38 and then for the System i. He led the teams responsible for Data Base, Commit, SMAPP, and Journal. His interests involve future performance and recovery improvements affecting journaling and IPL duration. He and the team he led have authored a popular IBM Redbook entitled "Striving for Optimal Journal Performance", numerous magazine articles regarding both Ragged SWA and Remote Journaling, plus over a dozen journal related TECHNOTES which address popular journal questions. These TECHNOTES can be accessed from IBM's redbooks website. |
|
High Availability on a Shoestring - - It’s Easier Than You Think!
If you thought purchasing an expensive high availability solution was the only game in town, think again. Yes, some shops clearly need a full-blown instant by instant replication approach so as to feel safe and assure a very rapid fail-over. Others, however, may be able to get by with a less expensive home-grown periodic refresh technique. This session will outline that less strenuous approach as it applies to the critical objects that can be journaled. If your budget is limited and as a consequence you thought no HA approach was feasible for your shop, come listen in as we explore other alternatives. |
A Better Save Strategy for your Database (Ragged Save While Active)
If you've got a tight save window, hate to kick off your end users, prefer to stay up 24x7, yet still need to perform nightly saves, i5/OS might have just what the doctor ordered: an enhanced and so-called "Ragged" Save While Active.
Traditional Save While Active has sometimes been called "Save While Not Too Active". Has that been your experience? If so, there's a new kid on the block and he's ready for the rough and tumble of true 24x7 operation. This new technique is especially useful in environments where your applications make heavy use of commitment control. Many of the popular ERP packages fall into that fold. So do most applications built with SQL.
In this session we'll show you the new SAVLIB parameters, explain how Ragged SWA differs from Classic Save While Active, and learn a thing or two about how to properly recover such a Ragged object when you get to the warm site.
By the end of this session, attendees will be able to:
- Identify the new parameters on the save commands.
- Grasp the benefits provided by the so-called “ragged” approach.
- Identify the proper recovery steps.
|
i5/OS - - A High Level Journal Fly-By
Journaling is at the heart of nearly all high availability approaches. Have you always wondered how journaling works and how best to employ this function? This session provides an introduction to i5/OS journaling and its functions.
This provides a good introduction for those who want to attend more in-depth journal sessions later in the week.
We'll introduce you to the objects, talk about hot site practices, show you how Journaling can affect your back-up strategy, show you how it can be a useful debugging aid, illustrate the role it plays as "better plumbing" in a high availability scheme, introduce many of the commands, suggest some optimal hardware choices that tend to be journal-friendly, and finish off with some performance tuning suggestions.
In this session, attendees will learn:
- When and where to use journal functions.
- Benefits to expect.
- Some of the major factors which affect performance.
- How to make applications more journal friendly.
|
A Smorgasbord of New Journal Choices
Come hear about the new features recently introduced for High Availability, improved performance, and faster recovery using Journal and Commit.
Topics include:
- A more auditor friendly space minimization approach.
- A kinder, gentler commit for improved performance.
- How altered command defaults can make life easier for journal users.
- Capacity increases where, "The sky's the limit" for journal users.
- Coaxing objects into starting out life in a journaled state.
- Data Queues: second class objects no more.
- Taming the mysterious journal recovery ratio tiger.
- Tell your EVI's to "come on down" and join the SMAPP party.
- Making your database cross reference file more robust so that you can bid reclaim storage adieu.
|
V6R1 Journal Choices to Get Your System Humming Like a Well Tuned Engine
V6R1 is chock full of new journal/recovery goodies. Many are directed at shops which are serious about achieving improved High Availability. If you're a serious journal user and intend to embrace V6R1 within the next 12 months, this session was designed with you in mind.
This session will show you how to respond, resolve and understand the following scenarios:
- Assuring maximum bandwidth and more robust resiliency for your remote journal connection.
- Using a new CRC-check to detect the presence of garbled TCP/IP transmissions and lost packets that might be influencing for high availability solution.
- Getting a better handle on that subtle, normally-hidden journaling behavior known as SMAPP so that you can tame that tiger.
- How to more easily reconfigure your journal environment on the fly in order to help balance the load across multiple journals (without having to close your files to do so).
- How to eliminate potentially "stale" data lingering in main memory if you stepped up to the journal caching option.
- Assuring all newly created files automatically fall under the journal umbrella so that no file goes unprotected.
- A brand new way to estimate journal overhead before putting your toe in the water (aka: Pseudo Journaling).
|
|
| |
| |
