------------------------------------
The scripts capture the following data sets:
------------------------------------
> Top Events - AAS CPU, latency, wait class
> SYSSTAT - Executes/sec, commits/sec, rollbacks/sec, logons/sec
> Memory - physical memory, PGA, SGA
> IO - IOPS breakdown, MB/s
> CPU - Load Average, NUM_CPUs,
> Storage - total storage size, per tablespace size
> Services - distribution of workload/modules
> Top SQL - PIOs, LIOs, modules, SQL type, SQL_ID, PX
> ASH data - detailed session activity (1sec and 10sec intervals)
------------------------------------
STEP 1: COLLECT AWR DATA (to be done by the customer)
------------------------------------
Unzip the scripts on a directory, for example: /home/oracle/dba/run_awr-quickextract
On the server, as SYSDBA run the following script on each database
if non-CDB environment: @run_all.sql
if CDB environment: @run_all_cdb.sql
The data collection end result will all be CSV files which will be consolidated on STEP 2
------------------------------------
STEP 2: CONSOLIDATE ALL FILES (to be done by the customer)
------------------------------------
After running the scripts consolidate all files in one zip file.
mkdir csvfiles_$HOSTNAME
mv *csv csvfiles_$HOSTNAME/
zip -r csvfiles_$HOSTNAME.zip csvfiles_$HOSTNAME
------------------------------------
STEP 3: EXTRACT & PACKAGE CSV FILES (to be done by the consultant/troubleshooter)
------------------------------------
- This section should be done on your laptop or another server
- The packaged CSV files are the files used for sizing and workload characterization
3a) Put all ZIP files in one folder, then extract the CSV files with the following commands:
mkdir zipfiles
for i in *.zip; do unzip $i -d zipfiles; done
cd zipfiles
3b) Package all CSV files
sh run_final_csv
zip awrcsvfiles.zip *txt