OID/Linux and Netsacpe Java SDK don't play nice

Here's a brief description of the problem. We are developing with the Netscape LDAP SDK for Java connecting to OID on Linux and deploying on Solaris. The problem I'm discussing here doesn't happen on Solaris. All access to the LDAP server is being funelled through one connection (please don't ask, I didn't design it).
Lets call an add, modify, delete or read an "action," for the sake of convenience. When an ldapjdk.jar client connects to an OID server, it can perform 32764 "actions." On the 32765th, it blocks forever waiting for OID server to return. Ok, the obvious solution is to implement a timeout and try again. Did that and it doesn't work. Ok, lets try dropping the connection at a watermark (say 30000 actions) and establishing a new one. Nope, still doesn't work. At the 32765 actions per client machine mark, it blocks forever.
Dumping the threads on the client side gives us something like the following:
main:
[1] java.lang.Object.wait (native method)
[2] java.lang.Object.wait (Object.java:424)
[3] netscape.ldap.LDAPMessageQueue.waitForMessage (LDAPMessageQueue.java:179)
[4] netscape.ldap.LDAPMessageQueue.nextMessage (LDAPMessageQueue.java:101)
[5] netscape.ldap.LDAPSearchListener.nextMessage (LDAPSearchListener.java:86)
[6] netscape.ldap.LDAPConnection.search (LDAPConnection.java:2584)
[7] netscape.ldap.LDAPConnection.read (LDAPConnection.java:2104)
[8] netscape.ldap.LDAPConnection.read (LDAPConnection.java:2098)
[9] ReadLockTest.run_test (ReadLockTest.java:96)
[10] ReadLockTest.run (ReadLockTest.java:49)
[11] ReadLockTest.main (ReadLockTest.java:25)
LDAPConnThread chakotay:389:
[1] java.net.SocketInputStream.socketRead (native method)
[2] java.net.SocketInputStream.read (SocketInputStream.java:90)
[3] java.io.BufferedInputStream.fill (BufferedInputStream.java:190)
[4] java.io.BufferedInputStream.read (BufferedInputStream.java:208)
[5] netscape.ldap.ber.stream.BERElement.getElement (BERElement.java:101)
[6] netscape.ldap.LDAPConnThread.run (LDAPConnThread.java:420)
Not much help there. I turned on all debugging on the OID server and got massive amounts of data back (and the server slowed to a crawl. :-] ). They all show the choke happening after gsldfb_FetchBaseObjs, and before leaving gslfcfdRefreshRootDSE. The last request (the 32765th) has started, but
never completes. I can provide parts of the logs on request.
Has anyone else encountered this problem?
Thanks.
Glenn McAllister
SOMA Networks, Inc.

Hello Glen:
I have not seen this before. Sounds like you have done a thorough job troubleshooting this and this may be a bug. If you are able, I would like you to file a trouble report at our Oracle Support web site:
http://www.oracle.com/support/metalink/index.html
Thanks,
Jay
null

Similar Messages

MSI RS480M2-IL (MS7093) and Powercolor X800GTO 16 don't play nice?

I have a MSI RS480M2-IL (MS7093) mainboard. My mainboard's integrated graphics work perfectly fine, but after I plug in the PCIe graphics card, problems occur: screen flickering, slow PC response, CPU usage 100%. I have already disabled the internal video mode, set initial display first to PCIe and set AGP aperture size to 128MB in the BIOS. Please help. Now I'm using the onboard graphics but I would really like to use the PCIe graphics card.
FYI the PCIe graphics card I'm using is the Powercolor X800GTO 16 256MB GDDR3 256 bit 399.21/489.86. I haven't done any overclocking to the card.
Does my mainboard support Nvidia cards? I tried a Gigabyte 6800GS but there were stability issues.

I just cleared the CMOS and the card is working fine now.

JPA and SQL Server 2005's Identity Don't Play Nice

I am hoping someone with a little more senior knowledge of JPASQL Server 2005 can point me in the right direction involving SQL Server 2005.
What I am discovering is JPA and SQL Server 2005 do not playing nice together when you use the IDENTITY capability for
a primary key. Let me explain. "...when using the IDENTITY as the generator type, the value for the identity field MAY NOT be
available BEFORE the entity data is saved in the database because typically it is generated when a record is committed."
My problem is this.
I have two separate entities using another entity, for example a car has a wheel and a truck has a wheel. I can go through the steps
of persisting the car and the wheel with no problem. My problem begins when I try to
find the wheel (by wheel description) for the truck. The query finds the wheel, but does not return the primary key for wheel.
Therefore, using the PK inside a query generates this error:
java.lang.IllegalArgumentException: An instance of a null PK has been incorrectly provided for this find operation. What has been my appraoch to get the entity to commit to the database is to use the following sequence for the Entity Manager.
entityManager.persist(wheel);
entityManager.flush(); I have assumed to this point this action would persist my data because I have been able to retrieve the PK immidately after these steps.
However, multiple actions taking place within the same transaction, appears to render the persist and flush steps useless. My initial thought
is to wrap the wheel creation within a new transaction so when the method/transaction completes, the data would be in the database. This
has not worked.
I hope I have explained this clear for others to follow. I apologize in advance if I have not. Any suggestions are greatly appreciated.

I am hoping someone with a little more senior knowledge of JPASQL Server 2005 can point me in the right direction involving SQL Server 2005.
What I am discovering is JPA and SQL Server 2005 do not playing nice together when you use the IDENTITY capability for
a primary key. Let me explain. "...when using the IDENTITY as the generator type, the value for the identity field MAY NOT be
available BEFORE the entity data is saved in the database because typically it is generated when a record is committed."
My problem is this.
I have two separate entities using another entity, for example a car has a wheel and a truck has a wheel. I can go through the steps
of persisting the car and the wheel with no problem. My problem begins when I try to
find the wheel (by wheel description) for the truck. The query finds the wheel, but does not return the primary key for wheel.
Therefore, using the PK inside a query generates this error:
java.lang.IllegalArgumentException: An instance of a null PK has been incorrectly provided for this find operation. What has been my appraoch to get the entity to commit to the database is to use the following sequence for the Entity Manager.
entityManager.persist(wheel);
entityManager.flush(); I have assumed to this point this action would persist my data because I have been able to retrieve the PK immidately after these steps.
However, multiple actions taking place within the same transaction, appears to render the persist and flush steps useless. My initial thought
is to wrap the wheel creation within a new transaction so when the method/transaction completes, the data would be in the database. This
has not worked.
I hope I have explained this clear for others to follow. I apologize in advance if I have not. Any suggestions are greatly appreciated.

Downloading and Installing Java SDK for AIX

Could anyone please tell me where can i download Java SDk for AIX and how to install it before installing SAP?
Thanks in Advance

Hi Mega,
You can downloads Java SDK 1.4.2 from this location, just choose the OS,
https://sdlc5d.sun.com/ECom/EComActionServlet;jsessionid=CA558AF3F6842F61D26641D9A6853DCA
installation is very simple, just copy the downloaded file and execute by running <b>./<name of the file>.sh</b> at the command prompt.
Regards
Juan
Please reward with points if helpful

Itunes 10.6 on Lion and Quicktime 10.1 don´t play bought movies from I tunes store, any ideas why?

I have issues with Itunes 10.6 on Lion and Quicktime 10.1. They don´t play bought movies from I tunes store, any ideas why?

I found the same issue... I can play other movies but no purchased ones. I can get the sound but the picture is scrambled. Does anyone know how to uninstall and reinstall iTunes? maybe that will work.

ARD and Time Machine don't play nice together

Hi there. Running 10.5.4 Server on a headless Mac Pro 8-core Xeon.
Thought i'd see how well time machine works as an incremental backup solution for my server. I use ARD to access the server. When i enter time machine using ARD i get some buggy glitchy version of time machine with no timeline or clear indication of where the controls are. Utterly unusable.
Any tips as to how to get these two magical pieces of software to play nice together?
in community,
~joel

Ahhhh -- so that's what's going on!
I have a Dual G5 Desktop that I'm connecting to, and it has worked perfectly -- except tonight.
I added about 1.6GB of data, and Time Machine has been spinning away!
Has anyone suggested how to get around this?
Stephen

Time Machine & Apple Remote Desktop don't play nice together

Hi there. Running 10.5.4 Server on a headless Mac Pro 8-core Xeon.
Thought i'd see how well time machine works as an incremental backup solution for my server. I use ARD to access the server. When i enter time machine using ARD i get some buggy glitchy version of time machine with no timeline or clear indication of where the controls are. Utterly unusable.
Any tips as to how to get these two magical pieces of software to play nice together?
in community,
~joel

When attempting to access Time Machine on a headless Mac over Apple Remote Desktop, a common complaint is that the GUI (graphical user interface) does not appear, though it does function if you guess where the buttons are.
To force the screen to redraw, just drag any file/folder from your local desktop into the Remote Desktop window, as if you were copying the file to the remote machine. You don't have to release the icon, but the screen should redraw with the Time Machine controls now visible and functional. Not all of the visual effects work, but you can browse past archives and perform restores with ease.
Keywords: Time Machine, ARD, Apple Remote Desktop, Headless, no monitor, no display, Mac Mini, Server
Enjoy.

Windows 7, Forms Runtime 6i and Discoverer 10g don't play nice together

Hi -
I posted this under the Discoverer Forum, but it was suggested that this might be a better place.
Let's see if I can have this make sense:
We are testing 3 Oracle products on a Windows 7 64-bit machine.
1. Forms Runtime 6i
2. Discoverer 10g version 10.1.2.3
3. ODBC from Ora client 11
I can install everything; although I had to install Discoverer in Compatablilty Mode using XP SP2. The ODBC client requires that I change the environment varialbes and add a New Environment variable called TNS_ADMIN.
Forms 6i works and the ODBC work until I install Discoverer. After installing Discoverer, I get a TNS cannont resolve.......on the Forms Runtime.
I'm lost as to how to get the Forms 6i to work. If you need any more information, please ask and I'll try to answer.
Thanks for all your help.

OK - so I thought the original post was from May of 2011 and wrote up my entire response accordingly. Only at the very end did I see that the original post was from May of 2010. Hopefully my response will help others who are only now trying to work through this issue.
Although Developer 6i is not officially supported by Oracle on Windows 7, it can run. It is just a bit finnicky to figure out.
There are a few of other things to I'd like to mention.
#1 Because you are running on a 64-bit OS, and depending on the Oracle 6i patch level you are running on the machine (we use Patch 17), you will need to replace two DLL files in the <ORA6i_HOME_DIR>\bin folder with version from 1/2/2001 (Patch 4, I think)
1. NN60.DLL
2. NNB60.DLL
If you do not replce these files, you will get an "FORM ERROR FRM-41211 Integration Error: SSL failure running another product" if you try to run a report from within your forms application, or if you try to open a form in a new window. The newer versions of these files apparently had removed some 64-bit calls when 64-bit Windows XP flopped.
More information (and my original source for my solution) can be found here: http://halimdba.blogspot.com/2009/05/oracle-forms-or-reports-6i-in-windows.html
#2 Because you are running on a 64-bit OS, if you want to add the TNS_ADMIN variable to your registry, you should add it in two places. HKLM\Software\Oracle AND HKLM\Software\Wow6432Node\Oracle
#3 During installation, I have a batch file that performs a lot of the 6i install stuff, so I don't run in compatability mode. But what I do is open a command prompt using "Run As Administrator" then browse to the 6i installers and install from that admin environment. If you are creating packaged forms 6i applications, I would strongly suggest investing the time to figure out how to script the 6i install for you clients. It can be a pain, but it is certainly much better than having to remember a gazillion different steps and make sure that you selec the approrpiate packages every time. The certainly don't need the full 6i environment, only the runtime portions.
#4 When installing multiple versions of oracle client, we've always found it best to start with the oldest version first. Get it running and tested, then move to the next version. Test the new version, then the old, then change the old to work with the new. Rinse and repeat. Of course, this is not required and may be impossible (installing your 6i app on machines that have been running 10g for years where you don't want to have to reimage the machine), but that order is definitely helpful in determine those steps that must be taken to get all versions working at the same time.

Enigmail and GnuPG Don't Play Nice With Permissions

In Thunderbird, I've got Enigmail to handle my GPG shit. The problem is that Enigmail needs to read .gnupg/private-keys-v1.d
STRING is the random string in the filename, I don't know if that is sensitive. If anyone can tell me what that is, bonus kudos.
enigmail> /usr/bin/gpg --charset utf-8 --display-charset utf-8 --batch --no-tty --status-fd 2 --import /home/c0rtex/.gnupg/private-keys-v1.d/STRING.key
gpg: can't open '/home/c0rtex/.gnupg/private-keys-v1.d/STRING.key': Permission denied
Anyways, it works fine if I change the permissions of .gnupg
However, the Arch wiki states that the permissions for .gnupg should be 700 and the files be 600 (which they are now). Otherwise gpg spits out a warning message. I don't know how big a deal this really is, there isn't much to be gained if my encrypted email were comprimised. But I try to implement best security practices on principal.
I'm feel this has to be an easy solution, change enigmails access privilidges?

The way I read the table at that link, the driver is already installed in Leopard -- so there is a chance that what you downloaded isn't as good as what shipped with Leopard.
In fact, clicking on the yes** in the table yields these instructions:
+If your printer is listed as bundled on Mac OS X 10.5, follow these instructions to install the printer.+
+Do not install the driver found on the installation CD or do not download and install driver found on the web. Refer to instructions below on recovering driver from Mac OS X 10.5 CD.+
+Plug and play+
+* Power on you printer.+
+* Plug the USB cable to your printer and to your Mac.+
+* The OS will automatically create the printer queue.+
http://www.lexmark.com/lexmark/sequentialem/home/0,6959,2048165961160493508_0en,00.html
(That link should get you squared away.)

Chromium and sound don't play nice together

Arch64, Chromium from the repos.
Long story short, if I play a YouTube video or anything similar, suddenly I can't play ANY sounds elsewhere until I close the browser down again.
Example: listening to music via MPD. I open a YouTube video. It doesn't play. I stop my music and watch the video. Once the video's done, I can't play any music until I shut down Chromium.

tomk wrote:
Working fine here - mpg123 playing a tune, youtube (flash, not html5) in chromium.
IIRC, this is because the flash plugin defaults to 'old' OSS i.e. using alsa oss emulation. Unload and blacklist the snd-*-oss modules and try again.
This works for me. If I don't have these modules blacklisted then I am able to reproduce the same problem as OP has. With the modules blacklisted everything works as it should.

CS6 and 'ZoneAlarm firewall + antivirus' don't play well together

I downloaded CS6 and it worked fine, that is until I downloaded ZoneAlarm firewall + antivirus beta. What I do to make CS6 work is to get off the internet by disconecting my PC, turn off ZoneAlarm and then CS6 starts working again. I wonder if anyone has the same problem? And how to get around it. Thanks.

Mojocelo wrote:
…And how to get around it…
Just another one of the reasons I use Macs.

After Effects CC and Premiere Pro CC don't play well together on new Mac Pros

Would love some insight from folks on this. We have welcomed 2 new mac pros w dual D700's into our edit suites and have found an issue we need some help finding a work around for.
With Creative Cloud we're used to rendering a large amount of graphics in After Effects, and while that's underway, our editors will often go on and work in Premiere on the edit. Unfortunately, with the new Mac Pros After Effects render process seems to freeze up the moment we open Premiere and start to work. Any fixes for this? Special Memory allocation? Patch? I've heard the lack of Cuda might be the issue, but can't imagine Adobe would not support something that keeps their product running smoothly.
Please let us know and we are grateful and thank you in advance...
[email protected]

This forum is actually about the Cloud, not about using individual programs
Once your program downloads and installs with no errors, you need the program forums
If you start at the Forums Index http://forums.adobe.com/index.jspa
You will be able to select a forum for the specific Adobe product(s) you use
Click the "down arrow" symbol on the right (where it says ALL FORUMS) to open the drop down list and scroll
http://forums.adobe.com/community/premiere
or
http://forums.adobe.com/community/aftereffects_general_discussion

OS 10.6.5, mac mini, Front Row, and Bose V20 don;t play nice

Since updating to 10.6.5 my 2010 mac mini using HDMI output to the Bose system will play one ripped video, then won;t play another without locking up (stuttering without audio) and then if I repair privileges, turn everything off and back on, and hold my tongue just right, the movie that wouldn;t play works fine without a lot of rhyme or reason... any ideas? when the audio strats acting up front row may not even play music without backing all the way put and then going back in...
(the log on privilege repair is pretty long, whereas before 10.6.5 it was almost nil)...
Thought?

Since the first post I think I have narrowed the issue to some kind of handshake between the mini and the Bose on the HDMI port... didn;t have an issue like this with OS 10.6.4. Might have to resort to Mini-DVI to HDMI and send the audio optically...

Issue with Youtube and Flash Player, videos don't play after a while

When I go to Youtube I can watch videos, but after I've player 5 or 6, when I select the next one, I get a snowy screen with a message "The Adobe Flash Player is required for video playback. Get the latest Flash Player". Issues is I already have the latest Flash Player, and it was playing fine just a minute ago. But from then on I cannot play more videos until I restart the computer. I'm tired of reinstalling Flash every day.
Win 7 64 bits, IE 9 and Firefox (happens with both)
Flash Player 11.7.700.202
Apparently the browser at first recognizes that Flash is installed, but then it doesn't find it anymore. Not sure how that happens.

strangest thing... a few 'non-youtube' sites still work ok, but still complain 'no javascript' ..
so went to firefox (v17.0.4 ESR) 'tools, options, content' and yup, javascript UNTICKED!! :O :O my bad memory? something else??? {shrugs}
ticked it, now ok........ think im getting old?????? ROFL..

Flash Player 9 and IE don't play nice

If you have a Flash movie with external calls, it crashes
when you try to navigate to another web page. You will receive an
'Out of Memory' error. This seems to be a problem with the way
Flash Player 9 handles the unload Handler when the web page's
javascript onUnload event is fired.
I have experiened this with both IE6 and IE7 RC1.
Has anyone else experienced this problem?

I came accross the same problem.
As of today, the only source I could find that addresses the
issue is:
http://blog.deconcept.com/2006/05/18/flash-player-bug-streaming-content-innerhtml-ie/#comm ent-14544
but their solution doesn't appear to be working reliably enough. I
am still getting "Out of Memory at line:56" errors even when I use
the suggested workaround.
I hope that the next version of the Flash Player will include
a fix for this.

OID/Linux and Netsacpe Java SDK don't play nice

Similar Messages

Maybe you are looking for