APEAR - A tool for automating Informatica Performance Tuning

Overview of APEAR

APEAR is a UNIX bash shell based tool that is capable of analyzing Informatica PowerCenter session logs and generate reports containing suggestions to improve session performance. The tool is FREE and tested with Informatica PowerCenter 8.5/8.6.

What does this tool do?

APEAR analyzes the informatica session log files and calculates and derives crucial performance insights from the log. Next APEAR uses this insight to suggest solutions or improvement points to augment the session performance and displays the suggestions in a nifty looking report. All these are done with zero manual intervention.

How can APEAR automate Informatica performance tuning?

APEAR uses a lot of calculation and heuristics (e.g. statistical deviations from normal distribution to identify query performance issue, etc.) and utilizes the generally accepted best practice guideline to suggest solutions.

In which platform APEAR runs?

APEAR needs Linux/Unix platform with bash shell

How does an APEAR report look like?

Please see a sample APEAR report below

AUTOMATED PERFORMANCE REPORTING FOR INFORMATICA MAPPING: m_3NF_GEN_DATA_LOAD_MAP_FINAL
 

PARTITION POINT                        | STAGE            |BUSY%      |COMMENT
---------------------------------------|------------------|-----------|---------------------
[SQ_DWH_DATA_LOAD_MAP_TEMP]            |read              |4.599      |
[SQ_DWH_DATA_LOAD_MAP_TEMP]            |transformation    |94.31      |Possible Bottle-neck
[DWH_DATA_LOAD_MAP_SNDWHTEMP1,         |write             |11.18      |
 
LOOKUP         |DTM BUFF|DTM BLCK| DCACHE | ICACHE |ROW(S)    |QTIME |CTIME |TTIME |TP% |DATA MEM |INDX MEM |FLAG
---------------|--------|--------|--------|--------|----------|------|------|------|----|---------|---------|----
LKP_DWH_DUMMY1 |609824  |65536   |14917632|7459200 |30        |0     |0     |0     |-   |         |         |
LKP_DWH_DUMMY2 |609824  |65536   |14917632|7459200 |285175    |1     |1     |1     |2.85|         |         |
LKP_DWH_DUMMY3 |609824  |65536   |14917632|7459200 |9528      |0     |0     |0     |-   |         |         |
LKP_DWH_DUMMY4 |609824  |65536   |14917632|7459200 |4853279   |87    |33    |87    |.55 |310624256|85828400 |PQ
LKP_SUB_DUMMY5 |609824  |65536   |14917632|7459200 |31449562  |410   |231   |410   |.76 |505167872|897080800|P
LKP_DWH_DUMMY6 |609824  |65536   |14917632|7458000 |792928    |3     |2     |3     |2.64|19062784 |35604800 |P
LKP_DWH_DUMMY7 |609824  |65536   |14917632|7459200 |129239    |2     |1     |2     |.64 |18915328 |3698800  |DQ
LKP_DWH_DUMMY8 |609824  |65536   |14917632|7458000 |1629778   |6     |3     |6     |2.71|39165952 |73185200 |P
ULKP_DWH_DUMMY9|609824  |65536   |14917632|7459200 |258       |0     |0     |0     |-   |         |         |
LKP_DWH_DUMMY10|609824  |65536   |14917632|7458000 |1950112   |10    |6     |10    |1.95|46858240 |87564400 |P
---------------|--------|--------|--------|--------|----------|------|------|------|----|---------|---------|----

TP% : Relative through-put of the lookup (1 mil rows/sec equals TP%=10)
FLAG: U- Unsorted Input, Q- Query may need tuning, P- Disk paging for Data and Index, D- Paging for data
 
SUGGESTION(s)

1.  LKP_DWH_DUMMY4: Not enough memory, data and index cache may page out.
[4853279] rows(of avg. len 64.00) needs 310624256(data) + 85828400(index) bytes of memory.
But it has 14917632 and 7459200 bytes respectively. You can increase memory from Properties
2.  LKP_SUB_DUMMY5: Not enough memory, data and index cache may page out.
[31449562] rows(of avg. len 160.62) needs 5051678720(data) + 897080800(index) bytes of memory.
But it has 14917632 and 7459200 bytes respectively. You can increase memory from Properties
3.  LKP_DWH_DUMMY7: Not enough memory, data cache may page-out to disk.
[129239] rows (avg. row len 146.35) needs 18915328 bytes of memory but it has 14917632 bytes
You can increase memory from Properties settings
4.  Possible bottleneck at partition point [SQ_DWH_DATA_LOAD_MAP_TEMP]
5.  The lookup sql query of [LKP_DWH_DUMMY4] may need some tuning
6.  The lookup sql query of [LKP_DWH_DUMMY7] may need some tuning

How can I download APEAR?

APEAR source code is freely downloadable from this website. Since this is basically a shell script, the code can be directly executed in UNIX bash.

APEAR1.0 - Initial Release

How to run/use APEAR?

APEAR can be executed by following the below simple steps:

1. Download the latest version of APEAR from the download link given above
2. Place the downloaded zip file in your UNIX/Linux server, preferably in the same server where Informatica is installed
   
3. Unzip the zip file using "unzip -j " command. This will create a .bsh file in the working directory
   $> unzip -j APEAR.zip
   
4. Execute the .bsh file from the command prompt and pass the full-path and name of the session log file for the session that you want to tune. See below
   $> chmod +x APEAR.bsh
   $> ./APEAR.bsh /SessLogs/informatica-session-log-file-name.txt
   

If your session log is in binary formatted, you may use the '-c' command line switch to convert the session log.

./APEAR.bsh -c/informatica/server/bin/infacmd.sh /SessLogs/informatica-session-log-file-name.log.bin

Where Can I report bug?

Please mail me "admin [at] dwbiconcepts [dot] com" to report any bug that you encounter.