The flight was a breeze (compared to some people in North Carolina that did not leave until late in the day!), had a great ride in compliments of the volunteer army from RMOUG, checked in to the conference just fine, and then had a hiccup visiting the Which Wich next door that was not staffed for a lunch rush from several conferences at the same time! I ended up a few minutes late to Chaitanya's presentation because of that!!
1:00 PM - Chaitanya Geddam (@GeddamChaitanya) - Deep Dive: Extreme Performance, Resiliency, and Monitoring of Improved GoldenGate 12c
1:00 PM - Chaitanya Geddam (@GeddamChaitanya) - Deep Dive: Extreme Performance, Resiliency, and Monitoring of Improved GoldenGate 12c
I really liked this presentation because it brought me up to speed on a technology that I do not use, but one that we do have in our environment at work so I need to understand it a lot better than I currently do!
General GG architecture - He had very good graphics and explained it well!
RBA address - This is a new type of address to me, which is like an SCN in the DB.
2 new types of replicats with 11g, coordinated (understands barrier transactions) and integrated
GG doesn't use SCNs for initial extract
11.2CC vs 11.2IC
Parameter files, obey scripts, INCLUDE files, custom shell scripts
Service resiliency
Bundled Agents use v4, not v5
MOS notes: 273674.1, 1298562.1, 1448324.1
What I thought was interesting is there are lots of ways to monitor GG but no single unified direction for the product!
Plugins don't understand VIPs and if 1 node is down it doesn't understand other nodes are working!
How to measure for trend analysis? Heartbeat Table Monitoring found in a MOS note so it looks like I am going to be digging around MOS when I get back home.
Is the table DBA_COMPARISON_SCAN just for GG? Time to find out!
3:15 PM - Riyaj Shamsudeen (@RiyajShamsudeen)- Deep Dive: RAC 12c Optimization
It is clear from Riyaj's presentation that he knows, like REALLY REALLY knows, the Oracle internals (probably why he has www.orainternals.com huh?) and while at some points it was too technical for the audience, I enjoyed how in depth he was digging into the RAC transactions!! I do not know if I will be digging that far down, but at least now I know it is possible and maybe I will think of a way to use this information in the future!
LMS is the workhorse of RAC, but remember it talks to LGWR too so that is just as important to tune!
I saw him call a DBMS_ROWID package that I am not familiar with so I will be looking that up when I get home too!
Database and instance transaction commits are different! That is a key point!
DBMS_CACHEUTIL was another package that I need to research when I get home!
The importance of NOT looking at the top wait events in a RAC environment was the biggest take away for me, and it gives me more items to research when I go home! Now I have another technology goal to add, and that is just after a couple sessions today! Imagine how many I will have tomorrow after a really FULL day of the conference?!?
The importance of NOT looking at the top wait events in a RAC environment was the biggest take away for me, and it gives me more items to research when I go home! Now I have another technology goal to add, and that is just after a couple sessions today! Imagine how many I will have tomorrow after a really FULL day of the conference?!?
No comments:
Post a Comment