Problem in tracing

Hi all,
I am thinking to use SAP provided Logging API in my web dynpro Application. I wanted to know that how can i view all the trace messages(i.e in which file).
How we can we use log-configuration.xml. i have seen the help file but it doesn't discuse it in details.

Hi,
Perhaps this link can be helpful
<a href="http://help.sap.com/saphelp_nw04/helpdata/en/81/9e99adb28ca4419787331dfc5ae0da/frameset.htm">configuring Logging API</a>
It has a link to the dtd, that is quite extensive.
However, you don't need to understand that file. It;s just for automatic configuration. You can configure the logging API just by usage of the Visual Administrator, tab log configurator.
By default the trace file and the output of your logs will be located at :
\usr\sap\<sid>\<instance>\j2ee\cluster\server0\log\defaultTrace.X.trc
HTH
Peter

Similar Messages

Problem in tracing session using ORADEBUG utility?

I want to trace a session using oradebug utility.what i have done is
SQL>oradebug setospid 25807
SQL> oradebug set ospid 27441
unknown command beginning "oradebug s..." - rest of line ignored.
How i can race another session using oradebug utility.
Thank you...

The correct syntax is
oradebug setospid <Unix pid>
You can 'attach' to one process only.
You have two oradebug setospid commands issued, the latter one is wrong, but wouldn't have worked anyway.
The usual method to trace a session is
oradebug setospid ...
oradebug event 10046 trace name context forever, level 12
This should work.
Note you can accomplish the same using dbms_system.set_ev
Sybrand Bakker
Senior Oracle DBA

Tracing report execution Problem

I am trying to trace report execution which gives me some problems!!!
First i tried to use report builder (via tools --> trace option) to add tracing options and then when i executed the report using View--> Runtime Priview it does not produce the trace file. This works if i am using Live Previewer instead of Runtime Preview.
But what i need is the trace file of runtime Previewer as it is giving me some problem.
If i programatically do it using report triggers using SRW built-in packages then the report builder itself closes and says RWBLD60 has generated an Error. Since my report is very complex i need the help of report tracing to confirm the flow and execution. Please can somebody shed some light into this problem....
It will be great if anyone tell me how to do it!!!!!.
i am using Windows 2000 & Reports 6i (6.0.8.11.3). IS THIS A BUG IN REPORTS 6i????
Thanks
Mahesh Mathew

The first thing I would do if I were you would be to upgrade to the latest patch version (we're up to about 6.0.8.23 now).
If you're still having problems with tracing you could contact support, or workaround by testing the report with rwrun60 and adding the tracing options on the command line.
Regards,
Danny

(HP Share to Web)/(java 1.5 JRE) incompatibility problem?

A Java application developed with JDK 1.3, and which used JRE 1.3, ran fine, but upon migrating from Java 1.3 to Java 1.5 (i.e. compiled with JDK 1.5 and using JRE 1.5), the application started having problems running on a given machine, which would give the fatal error, "javaw.exe has generated errors and will be closed by windows." This error would occur after the application would start and when the user would start to interact with application. The problem was traced to the application HP Share-to-Web, which was running on the machine in question. This problem would happen on machines running Win2K and Win98, and the aforementioned application, HP Share-to-Web, but would not happen on XP. When the application, HP Share-to-Web, was uninstalled, the Java 1.5 application would work fine. The following log file concerning this error was written on the Win98 machine. Hey..if anyone has any idea what be a happening, or if someone has experienced this same problem, any response would be appreciated.
JAVAW caused an invalid page fault in
module AWT.DLL at 0167:6d0ea81e.
Registers:
EAX=00000000 CS=0167 EIP=6d0ea81e EFLGS=00010246
EBX=00000000 SS=016f ESP=0b76fa34 EBP=817858a0
ECX=00000001 DS=016f ESI=006f84c0 FS=120f
EDX=81780e9c ES=016f EDI=00000000 GS=0000
Bytes at CS:EIP:
8b 07 57 ff 50 58 ff 76 18 8b 07 57 ff 50 58 ff
Stack dump:
006c1b80 006f8500 006f84c0 6d0ea7d1 00000000 6d0ea924 00000001 0042f5d0
7ff38b7d 006f8500 0042f5d4 0042f5d8 00000000 7ff387cf 7ff380f8 00000002

AWT strongly suggests gui to me.
That often means video software or hardware problems which can also include fonts.
Does the particular software package grap the screen? And is it certified to work on the OS with the particular video hardware and software (including versions)?

Session problem in LSMW.

Hi,
        For uploading the FIXED ASSETS to system, i am using the Standard Batch input method with object 0160, method 0001 and program name RAALTD01.
     The problem is with the create batch input session step. This step is directly processing the data without allowing the session to be processed later.
   If the data is having error, it is creating session with those error records. But for the ones that are successfull we are not able to see the new no generated, which can cause problem in tracing the conversion process.
Any help on this will be greatly appreciated.
Thanks & Regards,
Sudheera K

Hi Naveen,
                     After double clicking on the create bach input session, i am getting a screen with title 'LEGACY DATA TRANSFER FI-AA'. And below it the selection screen is given for fields 1) input file name and 2) Error session name.
Below it 3 check boxes.
    1) Test run without DB update.
    2) Transaction run in foreground.
    3) Structure of transfer formats.
When i click execute by giving the converted file name and the error session name, I am getting a screen which shows the no of records read, no of updated and no of error. Below that it is showing the error, if any any error record exists and a statement that the error session is created.
And on the status message it is showing the last asset number generated. For example if 45, 46 and 47 are the assets generated now. It will show in the status message that 47 is created.
Regarding the debugging, i will do that and let you know if i had any success.
Thanks & Regards,
Sudheera K

Very strange on tracing function

Hello All,
       I have encountered a very strange problem on tracing funciton.
       If set the disable tracing on a sequence,I will get a wrong result(all failed),And if enable the fuction,The test result is correct.
      Why will get total different result?
      Thanks a lot.
BR

Thanks for your reply.
We used the HMI built by labview,You can see all the HV test are failed.Normally I didn't simulate any error on product and expect all passed
Then I tried to open the dynamic sequence for debug,And set all the subsequences of "H
IPOT_5_STEPS_3" (sequence call trace settting as "use current trace setting".
Execute the sequence with Teststand,Will get the sequence(HIPOT_5_STEPS_3 is passed)

Emacs problem in Terminal.app

When I edit a text file in the (text-only version) of emacs running in Terminal.app, what happens all the time is that characters in my file get overwritten by spaces (on the terminal only). Pressing Ctrl-L recentres and refreshes the display and it's OK - until it happens again, which is usually right away.
E.g. if I open the file and press down rightarrow, and hold it for autorepeat, the cursor will gradually move through the file, but as I scroll over certain characters, they will be overwritten by spaces on the screen.
The problem only happens in Terminal.app, not also in xterm.
It doesn't matter whether $(TERM) is set to xterm, xterm-color, vt100 etc. in Terminal.app.
I remember this working in previous versions of Mac OS X, but not in 10.4.5 (and as I recall it's not worked for me for a while).
Is this a bug? Does anybody else see this? Or is there some setting I might have inadvertently changed.

The problem was traced to a
stty -tabs
line in my .login file. It appears that emacs outputs tabs to move around the screen, and those were being expanded to spaces which overwrote my text. Vi apparently does not use tabs in the same way, which is why the problem was limited to emacs.
The reason the problem manifested itself in Terminal.app but not in xterm is that tcsh executed inside a new Terminal.app window executes .login, whereas one executed inside a new xterm window does not.

Enable tracing

hii..
i m using OAS10g,want to enable tracing through form level...ie trace a particular user.
i made the user to use the following link to load his application so that i can creat a trace file in my $ORACLE_HOME/forms/trace...
ie http://ip adddress:7778/forms/frmservlet?config=xyz&record=forms
but to my surprise no trace file was generated.
should i need to add some parameter in formsweb.cfg..
for enablin tracing at form level
i read some article and it stated
http://download.oracle.com/docs/cd/B14099_19/web.1012/b14032/tracing002.htm
In order to trace Forms Processes on Windows, the Process Manager Service needs to have the check box "Allow service to interact with the desktop" selected.
but i neva got any check box wen i went from my em console..
please help,its bit urgent
Regards
Fabian

Ya,
i have deployed a war file..rite i solved the problem of tracing...
now the problem i m facing is
I generated the trace file..since it is in binary format want 2 convert it into xml format
so i apply this command
java oracle.forms.diagnostics.Xlate datafile=forms_10524.trc xmlfile=myfile.xml
Exception in thread "main" java.lang.NoClassDefFoundError: oracle/forms/diagnostics/Xlate
i searched on google about this error...
And it stated include the name in class path..and library path
but i dont have this location at all???oracle/forms/diagnostics/Xlate???
Do i need any jar files for for running this command(xlate)..
please help itz bit urgent
Thanks in advance
Regards
Fabian

Problem in Req/Reply in DBAdapter

Hi.
I need help in this problem urgently.
I have developed a integration schema where there are involved 3 applications (using DB adapter). The first app is a Human Resource App, The second one is MailSender App, and the third one is a Client App (Email r stored).
The integration schema developed is the following. When a person is retired from the company the event Pub/Sub is fired from the HR App sending this message to the MailSender App. This MailSender App ask to the Client App the email of the client whose Sales Man was retired. The integration method is made with a Synchronous Req/Reply DB - DB adapter.
The problem is when the event is fired, the Sync Req/Reply blocks for the given x sec to wait, but this never returns the answer. But checking the log, this inform that the DB Agent put the answer back exactly x sec specified in the timeout. If the Req/Reply is set to wait for infinty this blocks forever having to call to the DBA to kill the blocking session.
The Sync Req/Reply is ok when fired localy from the MailSender App, but when this is launched from the HR App is when this event blocks for the x timeout seconds. I think this is a lock issue with the interconnect integration scheme using DB adapter.
Any suggestion, patch, workaround, tec. pls inform me
thx

Hi,
I too had this problem where my Sun Req/Resp integration model expected result from SAP system. The problem I traced to was that, The procedure call will wait for Xsecs as mentioned and exits. With in this Xsecs, SAP was not sending the dat. I increased the duration. Also another issue was that the return message was of Char type and this was appending a lot of blank space, due to which in Log file there was lot of balnk space. I did a transformation to trim the char and it worked.
I don't know wheather this will be the sort of answer you expect.

Hp laserjet M1530dnf MPF slow printing

The technical specification for this printer is a print speed of 26 ppm. Last night I attempted to print a 31 page PDF file. It printed nine pages in 22 minutes before I cancelled the print job. I have an HP Deskjet 6600 for color printing. I sent the same file, set for B&W, to that printer. It completed printing all 31 pages in less than six minutes. Both printers use a wireless connection. Printer drivers for both were updated on 12 Apr 15. All Microsoft updaters have been applied. The 6600 is about three years old. I've had the M1530dnf Laserjet for about 45 days. This is just the most egregious example of the Lasterjet's abominably slow printing. Even a five page Word document can take between 3-to-5 minutes. The device manager reports the printer is working properly, but that is obviously not the case.
Any thoughts on where to go next are appreciated. The printer is still under warrenty, so I can take the problem back to HP. If thast's my option, does anyone have as good phone number to use? I used the number on the support web site and went through three people and over half-hour on hold only to get disconnected. I'd like to streamline that process by contacting the right person on the first call.
tia
This question was solved.
View Solution.

As it turned out, the problem was traced to a needed firm ware update. HP did a remote control section, had the problem isolated very quickly, performed the correction, and all within a half-hour the printer was back where it should be,
Everybody gets kudus for helping; everything you suggested got me closer and closer. Completeing the steps you outlined helped me sound a lot more intelligent. When I spoke to HP, I had accounted for most of the possible variables. That way, the time with them was much more productive.

Podcast downloads as MPEG, but stored as AAC?

Hi there,
I have encountered a strange issue when downloading certain podcasts from iTunes. The Podcasts in question are available for free on iTunes Podcasts and are available from 'MINUS Connections'. I'm not sure if this is a problem with my iTunes setup, a bug in iTunes, or the way in which the Podcast has been encoded. Nevertheless, here is the issue -
I download a Podcast from Minus Connections i.e. Minus Connection June '09 - Hearthrob
The Podcast shows as being of MPEG format while it is downloading
Once the Podcast has downloaded, it now displays as AAC format.
The Podcast plays in iTunes OK. However, in my own iPhone app, which reads an audio file and converts it to Linear PCM, it fails to read \ convert it.
My suspicion is that the file cant be read, because I am attempting to read it as AAC, when in fact it is actually an MPEG file?
So for some reason, at some point, the file kind has changed from MPEG to AAC, which is incorrent i think?
Interestingly, if I right click on the file in iTunes, and select 'Create AAC Version', it allows me to create an AAC version, which can be played OK on my iPhone app. I dont think the actual audio file format cant be read, I think it cant be read because the format type changes to the wrong type.
Has anyone encountered anything similar?
Many thanks,
David

MPEG-4 (.m4a) is AAC. There's nothing odd in iTunes or QuickTime player about the file for the most recent episode. I don't have an iPhone so I can't say what the problem is: a similar problem was traced to the embedded image being too large for the iPhone, but it's 600x600 here and it doesn't look as if the Chapter image is any larger, so that presumably isn't the problem - however using the 'convert to AAC' option removed the Chapter image (the playlist) though not the embedded image, so that may be the issue. You could ask in the iPhone Forum.

Is anyone seeing the flashing globe on startup in Mavericks?

When I startup my MBA running Mavericks 10.9.2, I was geting a flashing globe and long delay before being presented with the Apple logo and, ultimatley, the login screen.
For what it's worth, now that Activity Monitor no longer shows a disk usage pie chart, I brought up Disk Utility and, to my amazement, saw, in addition to the legit Macintosh HD boot drive, an AdobeFlashPlayerInstaller .dmg file shown as a selectable startup disk! That disk image file was in downloads; I deleted it from that folder and emptied the trash. I rebooted. No flashing globe - problem solved. I post this becasue all of the discussion threads I could find on this issue were pre-Mavericks, in the 2007 - early 2013 time frame. Well, the issue exists and this is a way to solve it.
What did SJ say about about Flash? A "bag of hurt"? I agree.
Hope one or more of you find this useful.
Pioneer.

I had another version of this problem, ultimately traced to a faulty USB device (my Apple keyboard). After repairing permissions, running Disk Utility, and reinstalling Mavericks, the flashing globe persisted. However, after the reinstall I also got error messages about multiple kernel extension problems. Apple Support suggested USB triage—removing all USB devices and plugging them back in one at a time and restarting. With everything unplugged, startup was normal. The keyboard showed up immediately as the problem. Plugged in a spare keyboard and the flashing globe appeared no more.

Help me with my survy for my projeect

I am doing this survy for my professor- I am at Bethedsa community college. I am already behind on this survey- so I will appreciate if you guys can quickly look at the survey file- One lucky winner will get $50 gift coupon at amazon.com
We are building a better version of JVM which will allow profiling in production. This technology, if feasible and successful may be ported to commercial application servers such as WebLogic, WebSphere, JBoss and others, with the cooperation of J2EE vendors.
In this document, we will be calling this system- "DiagnoseNow".
The DiagnoseNow project is in research stage, and we want to adjust the goals of the project to better meet the requirements of the application administrator, designer and IT manager.
Your inputs will be tremendously appreciated.
I will appreciate your response- I need to get 20-30 responses-
before my professor will release my stipend... I am sure you were a
student once- So I will really really appreciate your inputs. You do
not have to be complementary or negative to the idea- just be
unbiased and honest...
PROFILING AND TRANSACTION TRACKING IN PRODUCTION
We are building a better version of JVM which will allow profiling
in production. This technology, if feasible and successful may be
ported to commercial application servers such as WebLogic,
WebSphere, JBoss and others, with the cooperation of J2EE vendors.
In this document, we will be calling this system- "DiagnoseNow".
The DiagnoseNow project is in research stage, and we want to adjust
the goals of the project to better meet the requirements of the
application administrator, designer and IT manager.
Your inputs will be tremendously appreciated.
We are giving away one $50 gift certificate at Amazon.com to one
respondent. We expect about 25-50 responses only- So your odds of
winning are high.
The features of this proposed JVM are best explained using an
example of an ecommerce site.
Ecommerce site scenario
Description of system
Consider a typical ecommerce site with the following features: User
Validation, Product Search, Shopping Cart Management and Checkout.
Each user that actually purchases or attempts to purchase a product
on the website is consider a "Tracking unit" or a "Session". This
new research JVM will allow the time spent in a
particular "Sessions" to be tracked at the method level.
System administrator can if he / she so wishes get a report with
the following columns:
Session ID, MethodName, Time spent in the method(inclusive of
called methods), Time spent in the method(excluding called methods)
Put simply, the goal is to support session level tracking and
profiling at the method level in the JVM itself.
This will allow faster debugging and application turnaround,
reducing application maintenance cost.
A production ecommerce site could have millions of users and many
more transactions per day. We are optimistic about supporting real
life production sites.
User Validation
1. login: This validates the user using user name and password.
2. logout: This logs the user off.
3. passwordVerification: Using a database, the password is
verified.
4. retrievalOfLostPassword: If the user has lost his/her
password the password is emailed to the user.
Product search
1. byname: Search product catalog by name.
2. byDescription: Search product catalog by Description
3. bySku: Search product catalog by SKU
4. byBrand: Search product catalog by brand
5. Shopping Cart
6. add: Adds items to the cart
7. remove: Removes Items from the cart.
8. update: Updates Items from the cart.
9. checkout: Checkout items from the cart, by asking for credit
card. .
10. retrievePastData: Retrieve user information from the past.
11. validateCreditCard: Validate credit Card by checking with
the bank.
12. sendConfirmationEmail: Send Email confirming the purchase.
13. sendConfirmationEmailShipping: Send Email confirming that
the product has shipped.
14. sendSurveyEmail: Send Email checking customer satisfaction.
What is a "Transaction" or a "Tracking Unit" or a "Session"?
Each user who searches the product catalog is considered a "Tracking
unit" or sessions. A session ends when the users leaves the website.
Transaction tracking/ reporting data at method level
Time spent by each user, in each and every method above is tracked
in two ways: Time spent in a method itself, Time spent in the method
and the called methods. This will be done for each and every
invocation of the method.
Email alerts
Emails can be generated whenever a particular method(eg.
sendSurveyEmail) takes longer than a specified threshold.
Analysis Reports
Analysis can be done on a variety of topics: Reasons for abandoning
shopping carts, Slow or underperforming parts of the application.
Thread dumps
If a particular method runs slower than a threshold value, then
optionally a thread dump may be taken and stored. In this way, even
if the slowdown occurs at midnight, the thread dump will still be
available.
Triggered heap profile
Similarly, if the system is running out of heap memory a heap
profiler may become active. Heap profile may also be taken
periodically, to allow analysis of heap growth.
The system is best explained using a few examples:
Slowdown in search-books
Problem: Analysis of sessions shows that a large number of
customers are browsing for books, but conversion to actual sales is
slow.
Analysis: The book database has grown in size, and the server
running the database has many more apps on it. Moving the database
to another machine solved the problem.
How and Why did DiagnoseNow help?
DiagnoseNow maintains history of all sessions. The history was
analyzed to show customers who were not converting to purchases- a
large proportion of them were searching for books, and abandoning
the site afterwards.
Slowdown in search-general
Problem: Search has slowed down for all products.
Analysis and Resolution: Analysis showed that all search methods
are taking 250% longer than normal. The slowdown happened after a
specific date- it was the date on which the OS was upgraded. Rolling
back the upgrade and reinstalling it properly resolved the issue.
How and Why did DiagnoseNow help?
DiagnoseNow maintains a baseline of past search method response
times. So when a slowdown happens, there is no time
wasted "apportioning blame" or debating whether a problem exists.
Credit card authorization- intermittent failure
Problem: Credit card authorization was failing intermittenty after
hours
Analysis/Resolution: Credit card authorization was failing
intermittently. The automatically generated thread dump showed a
faulty connection. The problem was traced to the credit card
authorization end, and was resolved.
How and Why did DiagnoseNow help?
Without the thread dump this problem could not have been solved, and
without DiagnoseNow the problem would not have been detected unless
the system administrator was able to take thread Dumps- for that to
happen the problem would have had to occur during normal business
hours. (The off hours system administrators are Unix technicians
with no app server or Java knowledge. )
Slowdown of overall system at midnight
Problem: The entire system slows down at midnight.
Analysis/Resolution: All methods show a slowdown around midnight,
degrading system performance.
How and Why did DiagnoseNow help?
DiagnoseNow detected the slowdown, and it was discovered that a
large admin program ran at midnight. Splitting up the work in the
admin program into 5 chunks, made the performance impact lot
smaller.
Slowdown at 9.05 pm Saturday/Sunday evening
Problem: The entire system performance degrades by 50% around 9.05
pm on Saturday/Sunday evening
Analysis/Resolution: The weekend cleaning crew was unplugging one
of the appservers, to plug in the vaccum cleaner.
How and Why did DiagnoseNow help?
DiagnoseNow allows method level historical tracking and system level
analysis. This allowed quick detection of an overall slowdown.
Without system level analysis, a variety of alerts may get
triggered, but the root cause may not be identified.
SURVEY
Your Name(Required to get the prize):
Your role: Developer/IT Administrator/Manager/ App server
administrator/Software QA
Email address(Required to get the prize):
Phone number(optional):
Place a tick mark against the question:
1. Reducing application maintenance cost is important to me.
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
2. Triggering a thread dump based on specific conditions is an
important feature
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
3. Tracking each and every session/transaction down to the
method level is an important feature
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
4. Rapidly localizing and diagnosing a problem is important to
me
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
5. Creating alerts based on overall transaction performance is
important to me
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
6. If DiagnoseNow is proven to be a stable system, then I will
be willing to pay 1500 dollars per CPU license fee
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
7. I can accept a CPU overhead of 8% for extensive monitoring
leading to reduced costs.
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
8. If DiagnoseNow is proven to be a stable system, then I will
be willing to pay 1500 dollars per CPU license fee(inclusive of 2
business day email support. )
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
9. If DiagnoseNow is proven to be a stable system, then I will
be willing to pay 7500 dollars per CPU license fee(inclusive of
support with 1 hour response time)
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree

Profiling in production is a pipedream.
With modern JVMs and processors, hundreds of thousands of methods will be invoked on multi-CPU machines in just one second. Tracking and analyzing this data is beyond the power of the fastest database.
To analyze one hour of data, 360 million methods will have to be handled for a 4 CPU Pentium/Xeon machine.
This is clearly impossible to do in real life, with current hardware. May be 5 to 10 years from now, things will be better- however, by then CPUs will be processing many more methods per second...
I think it is a nice ACADEMIC project!

Please help me with my survey...

I am doing this survy for my professor- I am at Bethedsa community college. (Currently, our JVM runs about 400% faster than commercial JVM, so this is definitely in research stage. )
I will appreciate your response- I need to get 20-30 responses- before my professor will release my stipend... I am sure you were a student once- So I will really really appreciate your inputs. You do not have to be complementary or negative to the idea- just be unbiased and honest...
We are building a better version of JVM which will allow profiling in production. This technology, if feasible and successful may be ported to commercial application servers such as WebLogic, WebSphere, JBoss and others, with the cooperation of J2EE vendors.
In this document, we will be calling this system- "DiagnoseNow".
The DiagnoseNow project is in research stage, and we want to adjust the goals of the project to better meet the requirements of the application administrator, designer and IT manager.
Your inputs will be tremendously appreciated.
PROFILING AND TRANSACTION TRACKING IN PRODUCTION
We are building a better version of JVM which will allow profiling in production. This technology, if feasible and successful may be ported to commercial application servers such as WebLogic, WebSphere, JBoss and others, with the cooperation of J2EE vendors.
In this document, we will be calling this system- "DiagnoseNow".
The DiagnoseNow project is in research stage, and we want to adjust the goals of the project to better meet the requirements of the application administrator, designer and IT manager.
Your inputs will be tremendously appreciated.
We are giving away one $50 gift certificate at Amazon.com to one respondent. We expect about 25-50 responses only- So your odds of winning are high.
The features of this proposed JVM are best explained using an example of an ecommerce site.
Ecommerce site scenario
Description of system
Consider a typical ecommerce site with the following features: User Validation, Product Search, Shopping Cart Management and Checkout.
Each user that actually purchases or attempts to purchase a product on the website is consider a "Tracking unit" or a "Session". This new research JVM will allow the time spent in a particular "Sessions" to be tracked at the method level.
System administrator can if he / she so wishes get a report with the following columns:
Session ID, MethodName, Time spent in the method(inclusive of called methods), Time spent in the method(excluding called methods)
Put simply, the goal is to support session level tracking and profiling at the method level in the JVM itself.
This will allow faster debugging and application turnaround, reducing application maintenance cost.
A production ecommerce site could have millions of users and many more transactions per day. We are optimistic about supporting real life production sites.
User Validation
1. login: This validates the user using user name and password.
2. logout: This logs the user off.
3. passwordVerification: Using a database, the password is
verified.
4. retrievalOfLostPassword: If the user has lost his/her
password the password is emailed to the user.
Product search
1. byname: Search product catalog by name.
2. byDescription: Search product catalog by Description
3. bySku: Search product catalog by SKU
4. byBrand: Search product catalog by brand
5. Shopping Cart
6. add: Adds items to the cart
7. remove: Removes Items from the cart.
8. update: Updates Items from the cart.
9. checkout: Checkout items from the cart, by asking for credit
card. .
10. retrievePastData: Retrieve user information from the past.
11. validateCreditCard: Validate credit Card by checking with
the bank.
12. sendConfirmationEmail: Send Email confirming the purchase.
13. sendConfirmationEmailShipping: Send Email confirming that
the product has shipped.
14. sendSurveyEmail: Send Email checking customer satisfaction.
What is a "Transaction" or a "Tracking Unit" or a "Session"?
Each user who searches the product catalog is considered a "Tracking unit" or sessions. A session ends when the users leaves the website.
Transaction tracking/ reporting data at method level
Time spent by each user, in each and every method above is tracked in two ways: Time spent in a method itself, Time spent in the method
and the called methods. This will be done for each and every invocation of the method.
Email alerts
Emails can be generated whenever a particular method(eg.
sendSurveyEmail) takes longer than a specified threshold.
Analysis Reports
Analysis can be done on a variety of topics: Reasons for abandoning shopping carts, Slow or underperforming parts of the application.
Thread dumps
If a particular method runs slower than a threshold value, then optionally a thread dump may be taken and stored. In this way, even if the slowdown occurs at midnight, the thread dump will still be available.
Triggered heap profile
Similarly, if the system is running out of heap memory a heap profiler may become active. Heap profile may also be taken periodically, to allow analysis of heap growth.
The system is best explained using a few examples:
Slowdown in search-books
Problem: Analysis of sessions shows that a large number of customers are browsing for books, but conversion to actual sales is slow.
Analysis: The book database has grown in size, and the server running the database has many more apps on it.
Moving the database to another machine solved the problem.
How and Why did DiagnoseNow help?
DiagnoseNow maintains history of all sessions. The history was analyzed to show customers who were not converting to purchases- a large proportion of them were searching for books, and abandoning the site afterwards.
Slowdown in search-general
Problem: Search has slowed down for all products.
Analysis and Resolution: Analysis showed that all search methods are taking 250% longer than normal. The slowdown happened after a specific date- it was the date on which the OS was upgraded. Rolling back the upgrade and reinstalling it properly resolved the issue.
How and Why did DiagnoseNow help?
DiagnoseNow maintains a baseline of past search method response times. So when a slowdown happens, there is no time wasted "apportioning blame" or debating whether a problem exists.
Credit card authorization- intermittent failure
Problem: Credit card authorization was failing intermittenty after hours
Analysis/Resolution: Credit card authorization was failing intermittently. The automatically generated thread dump showed a faulty connection. The problem was traced to the credit card authorization end, and was resolved.
How and Why did DiagnoseNow help?
Without the thread dump this problem could not have been solved, and without DiagnoseNow the problem would not have been detected unless the system administrator was able to take thread Dumps- for that to happen the problem would have had to occur during normal business
hours. (The off hours system administrators are Unix technicians with no app server or Java knowledge. )
Slowdown of overall system at midnight
Problem: The entire system slows down at midnight.
Analysis/Resolution: All methods show a slowdown around midnight, degrading system performance.
How and Why did DiagnoseNow help?
DiagnoseNow detected the slowdown, and it was discovered that a large admin program ran at midnight. Splitting up the work in the admin program into 5 chunks, made the performance impact lot smaller.
Slowdown at 9.05 pm Saturday/Sunday evening
Problem: The entire system performance degrades by 50% around 9.05 pm on Saturday/Sunday evening
Analysis/Resolution: The weekend cleaning crew was unplugging one of the appservers, to plug in the vaccum cleaner.
How and Why did DiagnoseNow help?
DiagnoseNow allows method level historical tracking and system level analysis. This allowed quick detection of an overall slowdown. Without system level analysis, a variety of alerts may get triggered, but the root cause may not be identified.
SURVEY
Your Name(Required to get the prize):
Your role: Developer/IT Administrator/Manager/ App server
administrator/Software QA
Email address(Required to get the prize):
Phone number(optional):
Place a tick mark against the question:
1. Reducing application maintenance cost is important to me.
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
2. Triggering a thread dump based on specific conditions is an
important feature
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
3. Tracking each and every session/transaction down to the
method level is an important feature
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
4. Rapidly localizing and diagnosing a problem is important to
me
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
5. Creating alerts based on overall transaction performance is
important to me
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
7. I can accept a CPU overhead of 8% for extensive monitoring leading to reduced costs.
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
8. If DiagnoseNow is proven to be a stable system, then I will be willing to pay 1500 dollars per CPU license fee(inclusive of 2
business day email support. )
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
9. If DiagnoseNow is proven to be a stable system, then I will be willing to pay 7500 dollars per CPU license fee(inclusive of
support with 1 hour response time)
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree

piss off forum spamming monkey

Help me with my survey..

I am doing this survy for my professor- I am at Bethedsa community college. (Currently, our JVM runs about 400% faster than commercial JVM, so this is definitely in research stage. )
I will appreciate your response- I need to get 20-30 responses- before my professor will release my stipend... I am sure you were a student once- So I will really really appreciate your inputs. You do not have to be complementary or negative to the idea- just be unbiased and honest...
We are building a better version of JVM which will allow profiling in production. This technology, if feasible and successful may be ported to commercial application servers such as WebLogic, WebSphere, JBoss and others, with the cooperation of J2EE vendors.
In this document, we will be calling this system- "DiagnoseNow".
The DiagnoseNow project is in research stage, and we want to adjust the goals of the project to better meet the requirements of the application administrator, designer and IT manager.
Your inputs will be tremendously appreciated.
PROFILING AND TRANSACTION TRACKING IN PRODUCTION
We are building a better version of JVM which will allow profiling in production. This technology, if feasible and successful may be ported to commercial application servers such as WebLogic, WebSphere, JBoss and others, with the cooperation of J2EE vendors.
In this document, we will be calling this system- "DiagnoseNow".
The DiagnoseNow project is in research stage, and we want to adjust the goals of the project to better meet the requirements of the application administrator, designer and IT manager.
Your inputs will be tremendously appreciated.
We are giving away one $50 gift certificate at Amazon.com to one respondent. We expect about 25-50 responses only- So your odds of winning are high.
The features of this proposed JVM are best explained using an example of an ecommerce site.
Ecommerce site scenario
Description of system
Consider a typical ecommerce site with the following features: User Validation, Product Search, Shopping Cart Management and Checkout.
Each user that actually purchases or attempts to purchase a product on the website is consider a "Tracking unit" or a "Session". This new research JVM will allow the time spent in a particular "Sessions" to be tracked at the method level.
System administrator can if he / she so wishes get a report with the following columns:
Session ID, MethodName, Time spent in the method(inclusive of called methods), Time spent in the method(excluding called methods)
Put simply, the goal is to support session level tracking and profiling at the method level in the JVM itself.
This will allow faster debugging and application turnaround, reducing application maintenance cost.
A production ecommerce site could have millions of users and many more transactions per day. We are optimistic about supporting real life production sites.
User Validation
1. login: This validates the user using user name and password.
2. logout: This logs the user off.
3. passwordVerification: Using a database, the password is
verified.
4. retrievalOfLostPassword: If the user has lost his/her
password the password is emailed to the user.
Product search
1. byname: Search product catalog by name.
2. byDescription: Search product catalog by Description
3. bySku: Search product catalog by SKU
4. byBrand: Search product catalog by brand
5. Shopping Cart
6. add: Adds items to the cart
7. remove: Removes Items from the cart.
8. update: Updates Items from the cart.
9. checkout: Checkout items from the cart, by asking for credit
card. .
10. retrievePastData: Retrieve user information from the past.
11. validateCreditCard: Validate credit Card by checking with
the bank.
12. sendConfirmationEmail: Send Email confirming the purchase.
13. sendConfirmationEmailShipping: Send Email confirming that
the product has shipped.
14. sendSurveyEmail: Send Email checking customer satisfaction.
What is a "Transaction" or a "Tracking Unit" or a "Session"?
Each user who searches the product catalog is considered a "Tracking unit" or sessions. A session ends when the users leaves the website.
Transaction tracking/ reporting data at method level
Time spent by each user, in each and every method above is tracked in two ways: Time spent in a method itself, Time spent in the method
and the called methods. This will be done for each and every invocation of the method.
Email alerts
Emails can be generated whenever a particular method(eg.
sendSurveyEmail) takes longer than a specified threshold.
Analysis Reports
Analysis can be done on a variety of topics: Reasons for abandoning shopping carts, Slow or underperforming parts of the application.
Thread dumps
If a particular method runs slower than a threshold value, then optionally a thread dump may be taken and stored. In this way, even if the slowdown occurs at midnight, the thread dump will still be available.
Triggered heap profile
Similarly, if the system is running out of heap memory a heap profiler may become active. Heap profile may also be taken periodically, to allow analysis of heap growth.
The system is best explained using a few examples:
Slowdown in search-books
Problem: Analysis of sessions shows that a large number of customers are browsing for books, but conversion to actual sales is slow.
Analysis: The book database has grown in size, and the server running the database has many more apps on it.
Moving the database to another machine solved the problem.
How and Why did DiagnoseNow help?
DiagnoseNow maintains history of all sessions. The history was analyzed to show customers who were not converting to purchases- a large proportion of them were searching for books, and abandoning the site afterwards.
Slowdown in search-general
Problem: Search has slowed down for all products.
Analysis and Resolution: Analysis showed that all search methods are taking 250% longer than normal. The slowdown happened after a specific date- it was the date on which the OS was upgraded. Rolling back the upgrade and reinstalling it properly resolved the issue.
How and Why did DiagnoseNow help?
DiagnoseNow maintains a baseline of past search method response times. So when a slowdown happens, there is no time wasted "apportioning blame" or debating whether a problem exists.
Credit card authorization- intermittent failure
Problem: Credit card authorization was failing intermittenty after hours
Analysis/Resolution: Credit card authorization was failing intermittently. The automatically generated thread dump showed a faulty connection. The problem was traced to the credit card authorization end, and was resolved.
How and Why did DiagnoseNow help?
Without the thread dump this problem could not have been solved, and without DiagnoseNow the problem would not have been detected unless the system administrator was able to take thread Dumps- for that to happen the problem would have had to occur during normal business
hours. (The off hours system administrators are Unix technicians with no app server or Java knowledge. )
Slowdown of overall system at midnight
Problem: The entire system slows down at midnight.
Analysis/Resolution: All methods show a slowdown around midnight, degrading system performance.
How and Why did DiagnoseNow help?
DiagnoseNow detected the slowdown, and it was discovered that a large admin program ran at midnight. Splitting up the work in the admin program into 5 chunks, made the performance impact lot smaller.
Slowdown at 9.05 pm Saturday/Sunday evening
Problem: The entire system performance degrades by 50% around 9.05 pm on Saturday/Sunday evening
Analysis/Resolution: The weekend cleaning crew was unplugging one of the appservers, to plug in the vaccum cleaner.
How and Why did DiagnoseNow help?
DiagnoseNow allows method level historical tracking and system level analysis. This allowed quick detection of an overall slowdown. Without system level analysis, a variety of alerts may get triggered, but the root cause may not be identified.
SURVEY
Your Name(Required to get the prize):
Your role: Developer/IT Administrator/Manager/ App server
administrator/Software QA
Email address(Required to get the prize):
Phone number(optional):
Place a tick mark against the question:
1. Reducing application maintenance cost is important to me.
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
2. Triggering a thread dump based on specific conditions is an
important feature
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
3. Tracking each and every session/transaction down to the
method level is an important feature
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
4. Rapidly localizing and diagnosing a problem is important to
me
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
5. Creating alerts based on overall transaction performance is
important to me
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
7. I can accept a CPU overhead of 8% for extensive monitoring leading to reduced costs.
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
8. If DiagnoseNow is proven to be a stable system, then I will be willing to pay 1500 dollars per CPU license fee(inclusive of 2
business day email support. )
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree
9. If DiagnoseNow is proven to be a stable system, then I will be willing to pay 7500 dollars per CPU license fee(inclusive of
support with 1 hour response time)
Disagree Somewhat disagree Neutral Somewhat Agree
Strongly Agree

piss off forum spamming monkey

Problem in tracing

Similar Messages

Maybe you are looking for