Exadata - Oracle Goldengate Lag more than 1 hour
Issue #:
GoldenGate lag obersved for 1 hour at target side.
Analysis #1:
DB IO is high ,
8bc49ftgm971k – 38 sessions are running.
From last 1 hour continuously more than 20 sessions
@core DBA:
Only delete statement running from GG, seems IO is high due to other SQL id (which I given) is causing problem in DB level.
Till now given for instance 4. Please find instance 3:
Core DBA Question:
Venu, thanks for sending the details. What is the execution history for that SQL on an hourly basis? Is it a usual SQL or something adhoc? I am looking at this purely from a GGADMIN replicat session perspective
Analysis #2:
If you look overall DB is facing IO issue. IO issue started from 10am today. I am still checking some more details
Given SQL found tuning suggestion today morning after 6:00 AM
IO is not reducing and still increasing. It is better to restart application to stop IO.
Thanks for stopping app module, Now I can see IO is slowly reducing.
Good news, IO came to normal. Can you start GG replicates now one by one?
Once everything is OK then we can start App modules.
Sqlid history
yesterday after 10:00PM the query execution plan become worst. And from last 5 hours it become more worst.
After DB patch yesterday, DB was restarted, and it took worst execution plan.
Before 10PM: 2081515204 --> this execution plan was good.
After 10PM: --> 2025339486 --> new execution plan took, and it was very bad. You can check below information.
Solution:
We created sql baseline with 2081515204, to work normally and removed old plans.
Comments
Post a Comment