JNI Works for application, but not when used over Tomcat...

I am having problems getting JNI to work when I try to run a Servlet that invokes native calls.
I have a small, simple java program, just to test the waters, that simple calls a small shared library. It works like magic when I run the program as a simple application.
However, when I turn it into a Servlet, I get the following runtime errors:
The first time I call the Servlet:
java.lang.ExceptionInInitializerError: java.security.AccessControlException: access denied (java.lang.RuntimePermission loadLibrary.native)
     at java.security.AccessControlContext.checkPermission(AccessControlContext.java:272)
     at java.security.AccessController.checkPermission(AccessController.java:399)
     at java.lang.SecurityManager.checkPermission(SecurityManager.java:545)
     at java.lang.SecurityManager.checkLink(SecurityManager.java:837)
     at java.lang.Runtime.loadLibrary0(Runtime.java:743)
     at java.lang.System.loadLibrary(System.java:820)
     at Native.(NativeApp.java:24)
     at NativeApp.service(NativeApp.java:15)
     at org.apache.tomcat.core.ServletWrapper.doService(ServletWrapper.java:404)
     at org.apache.tomcat.core.Handler.service(Handler.java:286)
     at org.apache.tomcat.core.ServletWrapper.service(ServletWrapper.java:372)
     at org.apache.tomcat.core.ContextManager.internalService(ContextManager.java:797)
     at org.apache.tomcat.core.ContextManager.service(ContextManager.java:743)
     at org.apache.tomcat.service.connector.Ajp12ConnectionHandler.processConnection(Ajp12ConnectionHandler.java:166)
     at org.apache.tomcat.service.TcpWorkerThread.runIt(PoolTcpEndpoint.java:416)
     at org.apache.tomcat.util.ThreadPool$ControlRunnable.run(ThreadPool.java:498)
     at java.lang.Thread.run(Thread.java:484)
On hitting reload (F5), I get this:
java.lang.NoClassDefFoundError
     at NativeApp.service(NativeApp.java:15)
     at org.apache.tomcat.core.ServletWrapper.doService(ServletWrapper.java:404)
     at org.apache.tomcat.core.Handler.service(Handler.java:286)
     at org.apache.tomcat.core.ServletWrapper.service(ServletWrapper.java:372)
     at org.apache.tomcat.core.ContextManager.internalService(ContextManager.java:797)
     at org.apache.tomcat.core.ContextManager.service(ContextManager.java:743)
     at org.apache.tomcat.service.connector.Ajp12ConnectionHandler.processConnection(Ajp12ConnectionHandler.java:166)
     at org.apache.tomcat.service.TcpWorkerThread.runIt(PoolTcpEndpoint.java:416)
     at org.apache.tomcat.util.ThreadPool$ControlRunnable.run(ThreadPool.java:498)
     at java.lang.Thread.run(Thread.java:484)
Before going and farther, here is the server specs:
Linux (cobalt, basically red hat)
java 1.3
recent version of tomcat
Now, just to be sure, I removed the native call from the servlet, and ran it again...no problem this time. So I know that it has to be the native call that is causing this.
I have to say that I am still a bit new to JNI (book is in the mail form amazon still, hehe)...any ideas would really be appreciated.
I have a feeling that this may have little or nothing to do with JNI...but it seems to be related in some way.
-Zach

Well, I have came to the same conclusion, that the problem that I am having is directly related to the security policy that Tomcat runs under.
However, I have not been able to figure out exactly what I need to change....
I have altered the tomcat.policy file in a number of ways, including (for testing purposes) allowing full permission on the site that has the JNI servlet on it.
I have also done the same for the folder containing the shared library, I don't know if mattered, but I tried to be sure.
At this point, I have ordered a number of books from Amazon that should (hopefully) have some answers, but it is still a bit confusing that after allowing all permissions (and rebooting) that it still failed to work.
Dose anyone know how that you would allow native methods to be called in a policy file?
Thanks again for any suggestions!
-Zach

Similar Messages

HT3387 When I use pages the languages that I mainly use are English and Hebrew. The spellchecker works for English but not with Hebrew. How can I add another language to the spellcheck?

When I use pages the languages that I mainly use are English and Hebrew. The spellchecker works for English but not with Hebrew. How can I add another language to the spellcheck?

http://m10lmac.blogspot.com/2011/06/extra-spell-checking-dictionaries-for.html

Password works for login but not for the Adobe CC Desktop application

Password works for login but not for the Adobe CC Desktop application

Please refer to http://helpx.adobe.com/creative-cloud/kb/troubleshoot-cc-installation-download.html
ou can try deleting OPM.DB from steps below and try again.
Mac OS: You can locate the OPM.db file in the \User\\Library\Application Support\Adobe\OOBE folder. To access the hidden user Library folder, see Access hidden user library files | Mac OS 10.7 Lion. - http://helpx.adobe.com/x-productkb/global/access-hidden-user-library-f iles.html
Windows: You can locate the OPM.db file in the :\Users\\AppData\Local\Adobe\OOBE folder. To view the hidden AppData folder, see Show hidden files, folders, filename extensions | Windows XP, Vista, Windows 7. -http://helpx.adobe.com/x-productkb/global/show-hidden-files-folders-ex tensions.html
3) update to the latest version of the Creative Cloud app.
Also we would like few more details:
When has this issue started occurring, was this after installation of any specific adobe product.
On which platform are you seeing this.
Also we would like to know if after above steps this issue occurs ever again in future.
You can refer to http://forums.adobe.com/thread/1239510
Regards,
Rajshree

Can firefox support "file:///" format to access local files (note 3 fw slashes) this works for IE but not firefox and is the format used in the "afghan war diary" from wikileaks

can firefox support "file:///" format to access local files (note 3 fw slashes) this works for IE but not firefox and is the format used in the "afghan war diary" from wikileaks

See http://kb.mozillazine.org/Links_to_local_pages_do_not_work

My sound works for music but not apps

My sound works for music but not apps

Try this - Reset the iPad by holding down on the Sleep and Home buttons at the same time for about 10-15 seconds until the Apple Logo appears - ignore the red slider - let go of the buttons. (This is equivalent to rebooting your computer.)
Check your settings. The iPads have a small switch on the right edge. It can be used as a rotation lock to keep the screen from automatically reorienting itself as you move around, but you need to have the tablet’s settings configured properly. That same switch, right above the volume buttons, can also be set to function instead as a mute button to silence certain types of audio.
If the switch is set to work as a mute button, you can change its purpose to “screen-rotation lock” by tapping the Settings icon on the home screen. On the Settings screen, tap General on the left side, and on the right side of the screen flick down to “Use Side Switch to.” Tap to select Lock Rotation or Mute to set the button’s function. Even if you set the side switch for your preferred use, you can still mute the Mini or lock the screen. Just double-click the Home button, and when the panel of apps appears along the bottom edge of the screen, flick the row from left to right with your finger. Tap the icon on the far left side of the row to either lock the iPad’s screen or mute the iPad’s alerts, notifications and sound effects. Music, podcasts and video are not muted unless you turn the volume all the way down.
iPhone: No sound or distorted sound from speaker
http://support.apple.com/kb/ts5180
iPhone: Can't hear through the receiver or speakers
http://support.apple.com/kb/ts1630
http://www.atreks.com/app-no-sound-on-ipad-4-%E2%80%93-what-to-do/
To solve some sound problems, just follow these simple steps
1. Go to Settings
2. Tap on General
3. Tap on Reset
4. Tap on Reset All Settings
 Cheers, Tom

Flash shockwave 11.5 works for youtube but not amazon instant video

Shockwave 11.5 r502 works for Youtube but not for Amazon Instant Video.

Hi,
It asked me if I wanted to install something to make the videos load faster -
Check Safari / Preferences - Extensions. Might be lurking there. If so, select the plugin then click Uninstall.
Check:
~/Library/Internet Plug-Ins (~) Tilde character represents your Home folder
Check:
/Library/Internet Plug-Ins
/Library/Application Support
/Library/Input Managers
And check your Safari downloads folder. If the file is still there, double click to open. Might be an uninstaller you can use.
Help here for troubleshooting Flash
Carolyn
Message was edited by: Carolyn

Wifi works at home but not when away from home

My Wifi works at home but not when away from home, what is wrong? I thought with a phone plan I didn't need to have a wifi connection.

You either need WiFi or Cellular (or both). If you have neither, you can't connect to the internet.
Most people use WiFi at home and Cellular while away from home, or connect to public WiFi networks where there are some.

Album Art works for most, but not all

I was reading previous forums on the subject, but no clear answer was given..
The album art shows up fine in itunes and my library and works for most of my music on my ipod, but how come some album's art does appear? I have it set manually and select 'Get Info' then 'Artwork' and add it from there, then drag in the finished product to my ipod. I've been doing that and it works fine for most of the albums, but every so often, the album art will not appear on my ipod. Am I doing something wrong? I take them all from the same source (amazon.com) but it makes no sense to me how some randomly don't appear.
I tried deleting the artwork and re-adding it, but nothing happened again. If you could answer why it works for most but not all or how I can get the album art to appear for the ones it will not by re-doing something, it would be appreciated. thanks.

I encountered the same problem in the beginning. When I compared the size of the original images, the working ones seemed to be smaller in general, so I reduced the size of all covers I intended to upload on the pod to 300x300 pixels.
This seemed to work - for 2 days. I bought the pod on friday, had pictures on sunday, but on Monday (after the first time the pod was fully discharged) none were left left.
So I ended searching this forum for related problems...
And decided to join my problem to this entry.

UIImpersonator tests work in FlashBuilder, but not when run from Ant

I'm in the process of converting a Flex 3 project to use FlexUnit4 tests.
Everything compiles in Flashbuilder 4.5 and runs nicely. I've converted all the tests
to use flexunit 4, which highlighted a few issues, but nothing too major, and all the
tests now pass when run in Flashbuilder.
So... now, I'm trying to convert the CI build to use the 4.5 SDK and flexunit4. So far,
so good. Everything runs and I get a nicely formatted JUnit report. Unfortunately, though,
any test that uses the UIImpersonator fails with an async method timeout, suggesting that
the CreationComplete event hasn't fired.
Here's a typical example of one of my UI test cases:
          [Test(async, ui)]
          public function testAvailableProductsSetupAdminForSell() : void {
               var view:OrderBasketView = new OrderBasketView();
               var user:CfxUser = new CfxUser();
               user.admin = true;
               ModelLocator.instance.userDetails = user;
               view.buyOrSell = OrderType.SELL;
               helper.createComponentAndAddListener(view, this, availableProductSetupAdminForSellCreationComplete);
          private function availableProductSetupAdminForSellCreationComplete(event:Event, view:OrderBasketView) : void {
               Assert.assertTrue(view.availableProductTypes.contains(ProductType.PRODUCT_1));
               Assert.assertTrue(view.availableProductTypes.contains(ProductType.PRODUCT_2));
               Assert.assertTrue(view.availableProductTypes.contains(ProductType.PRODUCT_3));
and helper.createComponentAndAddListener looks like this:
        public function createComponentAndAddListener(view:UIComponent, testCase:Object, creationComplete:Function) : void {
               _view = view;
               _view.addEventListener(FlexEvent.CREATION_COMPLETE, Async.asyncHandler(testCase, creationComplete, 4000, _view));
               UIImpersonator.addChild(_view);
If I use FlexGlobals.topLevelApplication.parent.addChild(_view) in place of UIImpersonator.addChild(_view), all the tests pass.
I was wondering if it's a function of the fact that my helper class has no metadata that indicates it's a ui test, but that wouldn't explain why it works in FlashBuilder and not when run with ant.
I also wondered if it was a function of running with headless server set to true, but when I changed it to false the same thing happened.
My environment is:
ubuntu 11.04
ant 1.7.1
Sun jdk 1.6.0_26
Flex SDK 4.5.1.21328
FlexUnit 4.1.0-8-4.1.0.16076
I'm using the auto-generated TestRunner.mxml
Any thoughts, anyone? Now that I can get it to work by adding the UI components to the topLevelApplication, at least I can make progress, but I'd like to get to the bottom of the problem, because that shouldn't be necessary.
Thanks in advance,
-Chrisl

Changing to the topLevelApplication did not work. What's funny is that it then failed on a completely unrelated test by hanging and never returning... I'm thinking there must be something else that is going on here, but it's not clear what... :-/ Here is an example of my setup/teardown and a test that work great in the UI but not in CI...
[Before(async, ui)]
public function setUp():void
    _fromToList = new FromToList();
    _fromToList.setStyle('skinClass', FromToListSkin);
    Async.proceedOnEvent(this, _fromToList, FlexEvent.CREATION_COMPLETE, 1000);
    FlexGlobals.topLevelApplication.parent.addChild(_fromToList);
    // UIImpersonator.addChild(_fromToList);
[After(ui)]
public function tearDown():void
    FlexGlobals.topLevelApplication.parent.removeChild(_fromToList);
    // UIImpersonator.removeChild(_fromToList);
    _fromToList = null;
[Test(async)]
public function should_remove_selected_item_in_to_list_to_from_list():void
    _fromToList.fromArrayList = _dpArray;
    _fromToList.toArrayList = _toDpArray;
    var sequence:SequenceRunner = new SequenceRunner(this);
    sequence.addStep(new SequenceSetter(_fromToList.toList, { selectedItem: _toDpArray[1]}));
    sequence.addStep(new SequenceWaiter(_fromToList.toList, FlexEvent.VALUE_COMMIT, 100));
    sequence.addStep(new SequenceCaller(_fromToList, _fromToList.remove));
    sequence.addStep(new SequenceWaiter(_fromToList, FromToListChangeEvent.FROM_TO_LIST_CHANGE_EVENT, 100));
    sequence.addAssertHandler(handleListHasChangedThenRemoveEvent, {});
    sequence.run();
private function handleListHasChangedThenRemoveEvent(event:FromToListChangeEvent, passThruData:Object):void
    assertThat(_fromToList.toArrayList.length, equalTo(_toDpArray.length - 1));
    assertThat(_fromToList.fromArrayList.length, equalTo(_dpArray.length - _toDpArray.length + 1));
    assertTrue(_fromToList.fromArrayList.contains(_toDpArray[1]));
So you can see that I'm using Sequences to manage waiting for stuff to get updated in the background, and basically testing when I select an item in a list and act on it that it updates the model like I expect. Again, works GREAT in the UI Runner.
I'm open to suggestions...

Dynamic Parameter List works on desktop but not when run on Crystal Server

Hi,
I have a report, and the database command query takes one parameter ({?Year}. There is a second parameter used for a record select which is a dynamic paramenter list (multiple select, required). When running the report from Crystal Reports 2008 on my desktop, it first prompts for a year, and then once that is entered, it will come back with the list (of Companies) to select from. It works very smootly and as desired.
But when I load this report to Crystal Reports 2008 Server, it will ask for the year (which is a static list), and this is good, but then it comes back with an empty dynamic select list on the next screen.
Are there any special caveats that I need to be aware of in regards to dynamic lists when running from the CR server?
Thanks!

Hi Pat,
This should work but a couple of things to check:
1) Are the two parameters linked to the same database? If the parameters are linked to fields from different tables in different databases then you need to make sure you set the Database Configuration and have the report log on to both databases.
2) Are Crystal Reports and Crystal Server the same version? I had problems with parameters after we upgraded to 2008. We had to install the latest service packs for both Crystal 2008 and the Enterprise Server. Then we installed Crystal 2008 on the same machine the Enterprise Server was on. Crystal Reports and the Enterprise Server share some common files. Some of those files were updated for Crystal but not for the Enterprise Server yet. Once we synched them our reports have run fine.
Good luck,
Brian

Tiger Server: Default MySQL works for phpMyAdmin but not command line ? ? ?

Greetings all,
I followed the excellent instructions at
http://discussions.apple.com/thread.jspa?threadID=132783&tstart=0
on upgrading PHP to version 5 and getting (the preinstalled) MySQL setup with phpMyAdmin
I can connect through the phpMyAdmin, create new dbs, etc. Problem is that now I CAN'T connect from the command line. (I originally set my new root password from the command line and that's the last time I could connect from there!
Now I get the dreaded ERROR 2002 when I attempt:
/usr/bin/mysql -h localhost -u root -p
If I use the IP address or hostname instead of "localhost" (after -h) I get a ERROR 2003. Dropping the -h and hostname from the line above still gets me a ERROR 2002.
My phpMyAdmin config is a connect type "TCP" not socket, which might be one reason it works and the command line doesn't, but I still don't know where I've gone wrong.
My /var/mysql/* is all owned by mysql and group www.
I'm finding lots of people asking about the ERROR 2002, but not finding any solutions (particularly for Tiger Server). I'd prefer to use the preinstalled MySQL to installing a new installation (the thread referenced above says upgrading or removing the MySQL installation is too hard to do) and there has GOT to be a reasonable explanation for why it works with phpMyAdmin, but not the command line.
Thanks in advance for any replies!
Many Mac OS X (10.4.3)

You know what? I think when I started mysql_safe, I did it with sudo. Maybe that is my problem?
If so, what user do I want to start it as? I've seen one suggestion of doing this:
./bin/mysqld_safe --user=mysql &

SQL*Plus variable assignment works for 8i but not 9i

We have scripts that connect to each database on the box and perform database tasks nightly. The scripts first set the environment then connects to each database with SQL*Plus. This works for 8i but fails in the recently created 9i environment. (Unix Solaris environment)
Because the SQL*Plus connection appears in many areas in the scripts, we assign the following connection string to a variable:
SQLPLUS="/usr/oracle/product/9.2.0/bin/sqlplus -s 'xxx/xxxxxx@xxxx as sysdba'"
Export SQLPLUS
echo "$SQLPLUS"
(This echo out correctly: /usr/oracle/product/9.2.0/bin/sqlplus -s 'xxx/xxxxxx@xxxx as sysdba')
It fails when the script calls the assigned variable:
$SQLPLUS <<-EOF
with the following error:
Usage: SQLPLUS [ [<option>] [<logon>] [<start>] ]
where <option> ::= -H | -V | [ [-L] [-M <o>] [-R <n>] [-S] ]
<logon> ::= <username>[<password>][@<connect_string>] | / | /NOLOG
<start> ::= @<URI>|<filename>[.<ext>] [<parameter> ...]
"-H" displays the SQL*Plus version banner and usage syntax
"-V" displays the SQL*Plus version banner
"-L" attempts log on just once
"-M <o>" uses HTML markup options <o>
"-R <n>" uses restricted mode <n>
"-S" uses silent mode
However, if I remove the SQLPLUS variable assignment and changed all occurrence of
$SQLPLUS <<-EOF to
/usr/oracle/product/9.2.0/bin/sqlplus -s 'xxx/xxxxxx@xxxx as sysdba' <<-EOF
Then the script runs successfully. But this solution is cumbersome.
Any ideas as to how to have the script work with assigning the SQLPLUS variable????
Any help is appreciated.

I has an immediate suspicion it might be related to the issue
mentioned in
http://otn.oracle.com/support/tech/sql_plus/htdocs/sub_var2.html#2_7
but this proved wrong: the SP2-306 still occurs in the latest
SQL*Plus.
I wonder what version of 8i you had working? With an old SQL*Plus
8.1.7.0 my connection failed the same as in 9.2 and 10i.
My solution was to do:
SQLPLUS='sqlplus -s'
UNPW='/ as sysdba'
$SQLPLUS "$UNPW" <<EOF
EOFThis worked in 9.2.0.5, 10.1.0.2 and 8.1.7.0.
One common security risk on UNIX remains: putting the username and
password on the command line. On some systems a "ps" command will
show the password to any user. If OS authentication cannot be used
for connection, perhaps putting the username/password in the SQL
script may be more secure?
A final note is that in SQL*Plus 10g, no quotes are needed around
AS SYSDBA, i.e.
sqlplus / as sysdba works from the OS prompt, whereas in 9.2 you need to do
sqlplus "/ as sysdba"This makes a solution easy:
SQLPLUS="/usr/oracle/product/10.1.0/bin/sqlplus -s xxx/xxxxxx@xxxx as sysdba"-- CJ

SQL*Plus assignment works for 8i but not 9i

We have scripts that connect to each database on the box and perform database tasks nightly. The scripts first set the environment then connects to each database with SQL*Plus. This works for 8i but fails in the recently created 9i environment.
Because the SQL*Plus connection appears in many areas in the scripts, we assign the following connection string to a variable:
SQLPLUS="/usr/oracle/product/9.2.0/bin/sqlplus -s 'xxx/xxxxxx@xxxx as sysdba'"
Export SQLPLUS
echo "$SQLPLUS"
(This echo out correctly: /usr/oracle/product/9.2.0/bin/sqlplus -s 'xxx/xxxxxx@xxxx as sysdba')
It fails when the script calls the assigned variable:
$SQLPLUS <<-EOF
with the following error:
Usage: SQLPLUS [ [<option>] [<logon>] [<start>] ]
where <option> ::= -H | -V | [ [-L] [-M <o>] [-R <n>] [-S] ]
<logon> ::= <username>[<password>][@<connect_string>] | / | /NOLOG
<start> ::= @<URI>|<filename>[.<ext>] [<parameter> ...]
"-H" displays the SQL*Plus version banner and usage syntax
"-V" displays the SQL*Plus version banner
"-L" attempts log on just once
"-M <o>" uses HTML markup options <o>
"-R <n>" uses restricted mode <n>
"-S" uses silent mode
However, if I remove the SQLPLUS variable assignment and changed all occurrence of
$SQLPLUS <<-EOF to
/usr/oracle/product/9.2.0/bin/sqlplus -s 'xxx/xxxxxx@xxxx as sysdba' <<-EOF
Then the script runs successfully. But this solution is cumbersome.
Any ideas as to how to have the script work with assigning the SQLPLUS variable????
Any help is appreciated

I has an immediate suspicion it might be related to the issue
mentioned in
http://otn.oracle.com/support/tech/sql_plus/htdocs/sub_var2.html#2_7
but this proved wrong: the SP2-306 still occurs in the latest
SQL*Plus.
I wonder what version of 8i you had working? With an old SQL*Plus
8.1.7.0 my connection failed the same as in 9.2 and 10i.
My solution was to do:
SQLPLUS='sqlplus -s'
UNPW='/ as sysdba'
$SQLPLUS "$UNPW" <<EOF
EOFThis worked in 9.2.0.5, 10.1.0.2 and 8.1.7.0.
One common security risk on UNIX remains: putting the username and
password on the command line. On some systems a "ps" command will
show the password to any user. If OS authentication cannot be used
for connection, perhaps putting the username/password in the SQL
script may be more secure?
A final note is that in SQL*Plus 10g, no quotes are needed around
AS SYSDBA, i.e.
sqlplus / as sysdba works from the OS prompt, whereas in 9.2 you need to do
sqlplus "/ as sysdba"This makes a solution easy:
SQLPLUS="/usr/oracle/product/10.1.0/bin/sqlplus -s xxx/xxxxxx@xxxx as sysdba"-- CJ

Table of Contents page numbers works for PDF but not for other formats

Hi,
I followed exactly was indicated here regarding creation of TOC.
http://docs.oracle.com/cd/E10415_01/doc/bi.1013/e12187/T421739T481157.htm
The page numbers generated correctly for PDF but not for RTF, HTML and XLS.
(HTML and XLS are understandably not paginated in nature but why not in RTF?)
Tried to make the simplest example work, with 2 headings each on a separate page. Doesn't work on either MSWord 2003 or 2007.
What am I doing wrong?
Regards,
Jarell
Edited by: Jarell on Jan 11, 2012 8:08 PM

Hi Kavipriya,
What is your email address? I have attached a file in Oracle Metalink, under SR 3-5172030861. named test_toc.rtf
You may use any xml because the template does not use any dynamic binding of values.
The template is basically 2 pages, the TOC in the 1st page. First TOC item (Heading 1) is in Page 1, and Second TOC item is in Page 2.
So the end result should be (PDF result from BI Publisher):
First TOC Item .................. 1
Second TOC Item ............. 2
But for RTF it outputs:
First TOC Item .................. 0
Second TOC Item ............. 0
And after doing Print Preview (once or twice)
First TOC Item .................. Error in Bookmark
Second TOC Item ............. Error in Bookmark
Hope this is detailed enough to illustrate the problem.
Edited by: Jarell on Jan 17, 2012 10:24 PM

Express works for mac but not for pc's

my airport express works for my mac and iphone, but it won't work with my wife's and friends' pc's. the password works for me, but when they attempt to enter the password their computers ask for a network key instead of the password. please help!

I had successfully set up my Powerbook and PC to use Airport Express on a closed network. Been working fine for 5 weeks until tonight when neither machine could find the network. I reconnected the Mac but I can't get the PC hooked up again. Any ideas?
Charles

JNI Works for application, but not when used over Tomcat...

Similar Messages

Maybe you are looking for