Adapter queues and thread allocation

I have browsed through the forums for Adapter threads and message queue related issues but still dont have complete clarity on this topic.
Scenario-
Say i have only one file adapter interface in my XI box, an that file adapter is reading a 500 mb file and it has a pollling interval of 5mins, the file is not completely read within the first 5 mins.
Question-
Now will the thread allocated to this adapter be alive and keep reading this file or will it be released when the next polling interval starts?
So what happens when the polling interval is exceeded?
(Assume there is no parallel processing)
will a new thread be allocated to the adapter ?
If yes will it start processing the file or will it wait until the completion of the first thread ?
According to the New Queueing Scheme in XI 3.0 SP19 / 7.0 SP11
a sender file adaper in sync mode has one queue assigned to it and Send.maxConsumers=5, say.
Consider a situation where i have 6 file adapters all reading 50mb files,
Will one of the file adapter threads will be blocked because all the 5 threads would be in use simultaneously?
If a large file is being read and is being written in the message queue all the other threads would have to wait till the file is fully read and will be in the blocked state right?
Appreciate any help.
Thanx
Siva

Hi,
For each adapter there are four queues..
File_http://sap.com/xi/XI/SystemCall
File_http://sap.com/xi/XI/SystemRecv
File_http://sap.com/xi/XI/SystemRqst
File_http://sap.com/xi/XI/SystemSend
and each of the above queues will have threads assigned.(5 by default)
If these queue threads are all busy at atime then the message status will be set in Waiting Status...
Once a message is read via Call queue ,data will be sent to Adapter engine Messaging System via Recv queue...
Similarly in recieving side data will be received from IE via Rqst and send to adapter via Send queue..if target system is busy i.e adapter thread is keep running and there are 6 instances of the same type then message status will be Waiting/To Be Delivered while the message status of processing messages will be Delivering.
So in your case data will be read via Call queue and passed on the Recv queue..which result in availability of queue ..even if set interval elapses..
Hope this gives you an insight
Regards
Rajesh

Similar Messages

Problem with Queue and threads

Hello,
the following is the code :
public class Tester extends Thread
private String fname;
public String getFname()
   return fname;
public void setFname(String fname)
   this.fname = fname;
public Tester(String fname){
   super();
this.fname = fname;
}Heres another class
class N {
    Thread r = null;
    public void checkQ()
      Queue q = new ConcurrentLinkedQueue<N>();
      for(int i = 0; i < 13; i++)
         System.out.println("FOR I is "+i);
         r = new Tester("Name" + i);
         q.offer(r);
         processq(q);
    public void processq(Queue queueOfM)
       if(queueOfM.size() > 10)
          System.out.println("size of queueofM is "+queueOfM.size());
        for(int j=0; j < queueOfM.size();j++)
             System.out.println("J is "+j);
             this.r = (Thread)queueOfM.poll();
             this.r.start();
   }When run, the code prints :
FOR I is 0
FOR I is 1
FOR I is 2
FOR I is 3
FOR I is 4
FOR I is 5
FOR I is 6
FOR I is 7
FOR I is 8
FOR I is 9
FOR I is 10
size of queueofm is 11
J is 0
J is 1
J is 2
J is 3
J is 4
J is 5
FOR I is 11
FOR I is 12I was expecting the code to print the J till 11.
Could you please help me find where I went wrong ?

I think I may have been asking a wrong question.
class N {
    Thread r = null;
    public void checkQ()
      Queue q = new ConcurrentLinkedQueue<N>();
      for(int i = 0; i < 13; i++)
         System.out.println("FOR I is "+i);
         r = new Tester("Name" + i);
         q.offer(r);
         processq(q);
    public void processq(Queue queueOfM)
       if(queueOfM.size() > 10)
          System.out.println("size of queueofM is "+queueOfM.size());
        for(int j=0; j < queueOfM.size();j++)
             System.out.println("J is "+j);
             this.r = (Thread)queueOfM.poll();
             this.r.start();
          }The if in the processq method will come true when I is 11. so, once in the if, the for loop is executed. The j is checked till it is < 11 and then the value of j is being printed. So, as I was expecting J should print till atleast 10. However it is printing till 5 and quitting the loop, I dont know how or why, for which I was thinking if someone could show me why.

Most efficient way to constantly read, queue and parse multi-sized RS232 data (multi-threaded)

I've tried tackling this problem a few different ways, and figured it was time to get some others' advice. My system essentially works, although it looks like a hackjob and not entirely confident.
My RS-232 connection has the following properties/constraints:
-Will be getting unsolicited data at a high data rate. (ya thats subjective, but assume near constant at whatever baud rate its set at up to 115200)
-Different segment lengths of data receiving
-Two stop bytes (0x10 0x03), while start byte is 0x10. (Byte stuffing/packing implemented)
-There is a size byte within a packet (3rd one in), however currently relying on stop bytes only.
I have tried the ComCallback within CVI only to find that it is VERY slow at processing events compared to implementing it manually in its own thread. In addition, it can only trigger 1 stop bit, not 2. Triggering on size is sometimes okay, but I found that it was possible it would get triggered on only part of the data, the when its called the qlength be larger, and then sometimes I would only read part of a data packet, and half of my segment was still in the queue. And sometimes, I would get semaphore locks and lots of waiting, and ya it was a mess (hence you will see lots of CMT locks commented out)
I tried implementing a FIFO type queue (copied below), but I have very little experience in doing this, and may not be very efficient in the way I implemented and could definately use some advice in this area. Also perhaps in how thread safe I have everything.
I thought about a circular buffer, but since data can be different segment lengths, it kind of makes it difficult to cleanly wrap around and read. I think its still possible, just may require additional checks which I haven't seen done anywhere when google searching. (and made me think a FIFO queue was better).
So if anyone has any good suggestions or examples, that would be great. Using Labwindows 2010.
//in Main before GUI loaded
programRunning = 1;
CmtScheduleThreadPoolFunction (DEFAULT_THREAD_POOL_HANDLE, ComCallback, NULL, &funcID);
/* Function Used to parse Com data */
static int ComCallback (void *callbackData)
static int bufLen = 0;
int strLen,qLen=0;
unsigned char tempBuf[1024];
int packet_length = 0;
char temp_string[250];
unsigned char *start_ptr;
int i;
int start_offset;
while(programRunning) {
//First lets copy all data from com port to output buffer
if( com_open) qLen = GetInQLen (comPort);
if( qLen <= 0 ) {
ProcessSystemEvents();
} else {
if( qLen > 1024) qLen = 1024-bufLen; //set max length to read
strLen = ComRd (comPort, tempBuf, qLen);
//CmtGetLock(lock);
memcpy(readBuf+bufLen,tempBuf,strLen);
bufLen += strLen;
//CmtReleaseLock(lock);
//Try to read until we hit stop bytes, or until we think we have at least 1 or 2 packets to process
if(tempBuf[strLen-2] != 0x10 && tempBuf[strLen-1] != 0x03 && bufLen < 100 ) )
goto skip;
//ensure start pointer is at beginning command (0x10 0xAA) (in event we just started reading in middle of packet???)
i = 0;
while ( (readBuf[i] != 0x10) && (readBuf[i+1] != 0xAA) && (i < strLen) )
//start_ptr++;
i++;
start_offset = i;
parse_some_more:
start_ptr = readBuf + start_offset;
//lets try to do one packet at a time.
for(i = start_offset; i < bufLen-start_offset-1; i++)
if(readBuf[i] == 0x10)
if(readBuf[i+1] == 0x03)
i += 2;
break;
//trial of unpacking/unstuffing byte buffer
else if( readBuf[i+1] == 0x10 ) //two tens in a row, remove it
memmove( &readBuf[i],&readBuf[i+1],bufLen-i);
bufLen--;
//at this point, we should have a full packet. What if we don't....???
packet_length = i - start_offset;
//CmtGetLock(lock);
ParseResponse(start_ptr,packet_length);
PostPacketToOutputBuffer(start_ptr,packet_length);
if(start_ptr[start_offset] == 0x10 && start_ptr[start_offset+1] == 0xAA )
com_Send_Acknowledge(comPort);
start_offset = packet_length + start_offset;
if(start_offset < bufLen-1)
goto parse_some_more;
//For now, assume everything else we don't care about in buffer, however, probably shouldn't. Should it be bufLen -=start_offset; if so, need to handle partial data better; timeout??
bufLen = 0;
//CmtReleaseLock(lock);
skip:
return 0;
Thanks!

Hi ngay528,
I think there is a great example for you to use that comes with CVI which can be found by clicking Find Examples on the splash screen or Help >> Find Examples in your project. From there, click into Optimizing Applications >> Multithreading. In that folder there is a project that is called BuffNoDataLoss that shows how to create a thread safe queue and setup a producer/consumer type program. In this example the data is a random sine wave but could be adapted to your RS232 data. If you have any questions concerning this example please let me know but this should be a great starting point.
Patrick H | National Instruments | Software Engineer

Lookup API - asynchonous SOAP Adapter Queue

Hi,
I have a question about the queue handling, when you use the Lookup API of the SAP PI. ([Sap Help Lookup API|http://help.sap.com/saphelp_nw04/helpdata/en/cf/406642ea59c753e10000000a1550b0/frameset.htm])
When I call a synchronous Web Service via SOAP within a message mapping via the Lookup API, the synchronous web service is called via the "SOAP_http://sap.com/xi/XI/SystemSend" adapter queue. But "SOAP_http://sap.com/xi/XI/SystemSend" is the asynchronous SOAP sender adapter queue. The threads which handle the calls are not made for handling synchronous calls. If the synchonous web service has a high response time (10 seconds for example). The thread is blocked for this time.
Synchronous threads are not blocked that long.
Is there a way to configure which adapter queue has to be used by the lookup API?
Regards,
Torsten

Hello,
Regarding the error in the trace:
nested exception is: com.sap.engine.services.deploy.exceptions.ServerDeploymentException: [ERROR CODE DPL.DS.5070] Already started operation with application sap.com/com.sap.aii.adapter.soap.app in the cluster. No other operation with this application is allowed at the moment, because it is locked in the enqueue server or the lock from there has been manually removed. at com.sap.engine.services.ejb3.runtime.impl.DefaultContainerRepository.startApp(DefaultContainerRepository.java:236) at
please see note
1339258 PI Adapter applications locked during start up
You could try restarting the SOAP Adapter applications :
com.sap.aii.adapter.soap.app
com.sap.aii.adapter.soap.lib
from the JEE NWA
-> Operation Management
-> Systems
-> Start and Stop
-> Java EE Applications
http://wiki.sdn.sap.com/wiki/display/XI/start_stop_applications
Regards,
Sarah

Understanding asynchronous BPEL processes and threading

Hi everybody,
I have a question regarding asynchronous BPEL processes and threading.
I have an asynchronous BPEL process which delivers a message (picked up from a JMS topic) to a repository (via for instance an ICAN server). This transaction must be time-sensitive, i.e. first-in is first-out.
If I have several threads running concurrently I have no guarantee that the messages are delivered in the same order that they where generated. It seems that the suggested way around this problem is to use single-threading.
So here's the part that I do not understand:
An asynchronous request does not wait for a response. If I use single-threading then what happens to that single-thread if for some reason (for instance the ICAN server is not responding) the BPEL instance is dehydrated? Is the thread still allocated for waiting for the first response or does it go back and pick up the next thing on the topic? If the next thing on the topic is picked up then how can I be sure that the messages are delivered in the correct order? If the thread is still allocated to the single instance then does that not kindda go against the whole concept of an asynchronous process?
I hope some one can explain this to me.
Thank you in advance.
Kind regards,
Mathias
PS. This question is an attempt of a different angle on another forum thread: Asynchronuos BPEL - how to get the order of messages right?

Hi again,
A quick follow up question:
Is there any way to have several BPEL processes running on the same server with different settings for how many threads to use?
It seems that the only way to set number of threads used by BPEL is to use the configuration in the BPEL console? This is an domain setting - which means that all processes deployed to the server is effected the same way?
I am asking because I am facing a scenario where I need one BPEL process to read from a AQ queue (well actually a topic) with a single thread and another BPEL process - deployed to the same BPEL server - to read from the same AQ queue with multiple threads.
Does anyone have any idea of how to do this? Can it be done?
Any suggestions, hints or reference to any documentation would be very much appreciated.
Thank you in advance,
Aagaard

Reading Opaque data from jms queue and decoded in java embedding

Hi ,
Objective:Fetch text message from queue and print it from java embedding in BPEL
I am fetching text message from a jms queue using JMS adapter in BPEL.Then converting the opaque data(Base64 binary)to string using java embedding.My build got successful but I am getting runtime error at the time of deployment.
I have added fabric-runtime.jar file into SCA-INF\lib folder as well as project library directory.
<extensionActivity>
<bpelx:exec name="Java_Embedding1" version="1.5" language="java">
<bpelx:exec import="java.io.*"/>
<bpelx:exec import="oracle.soa.common.util.Base64Decoder"/>
<![CDATA[//Write your java code below e.g.
     System.out.println("Hello, World");
//   String input = ((Element)getVariableData("Receive1_Consume_Message_InputVariable")).getTextContent();
// String inputName = (String)getVariableData("input");
String input = (String)getVariableData("Receive1_Consume_Message_InputVariable");
//Receive1_Consume_Message_InputVariable.opaque
      System.out.println("input:"+input);
      addAuditTrailEntry(input);
      Base64Decoder Decoder = new Base64Decoder();
     try {
      String decoded = Base64Decoder.decode(input);
      setVariableData("Variable_opaque_decoded",decoded);
      addAuditTrailEntry(decoded);
catch(UnsupportedEncodingException uee)
uee.printStackTrace();
}]]>
</bpelx:exec>
</extensionActivity>
Error:
Error deploying archive sca_JMSToDBInsert_rev1.0.jar to partition "default" on server AdminServer [http://localhost:7001]
HTTP error code returned [500]
Error message from server:
There was an error deploying the composite on AdminServer: Error occurred during deployment of component: BPELProcess1 to service engine: implementation.bpel, for composite: JMSToDBInsert: ORABPEL-05250
Error deploying BPEL suitcase.
error while attempting to deploy the BPEL component file "C:\Oracle\Middleware\user_projects\domains\soa_domain\servers\AdminServer\dc\soa_0fb06e23-bac3-4753-aae9-0fc4b54b4dd1"; the exception reported is: java.lang.RuntimeException: failed to compile execlets of BPELProcess1
This error contained an exception thrown by the underlying deployment module.
Verify the exception trace in the log (with logging level set to debug mode).
Check server log for more details.
Error deploying archive sca_JMSToDBInsert_rev1.0.jar to partition "default" on server AdminServer [http://localhost:7001]
#### Deployment incomplete. ####
Error deploying archive file:/C:/SOA Suite/SOA_POC/Client2JMSapps/Client2JMSapps/JMSToDBInsert/deploy/sca_JMSToDBInsert_rev1.0.jar
(oracle.tip.tools.ide.fabric.deploy.common.SOARemoteDeployer)
Could you guys please help me regarding this?
Thanks & Regards,
Souvik
Edited by: Souvik Pal on Apr 16, 2012 9:40 PM

You probably want to try the SOA Suite forum for OSB questions.
SOA Suite
There is an excellent book also:
http://jeffdavies.org/
The documentation goes over the fundamentals, but the book will be better for step by step examples.

Message flow between advanced adapter engine and messaging system for ICO scenarios

Dear all,
I'm working on a PI 7.31 AEX box and created an ICO object for one interface with HTTP_AAE2IDoc scenario. The interface works quite well. When I looked in to message log of messages for this interface in PI monitor, I noticed that the internal XI message are put totally 5 times into the send queue and retrieved from send queue. In some steps I only see after retrieving a message from send queue and immediately put it back into send queue. Can someone please kindly explain the message flow between advanced adapter engine and messaging system and what exactly happens after each retrieving a message from send queue?
Below is an example of message log for one message:
12/08/2014 09:26:06.472    Information    MP: processing local module localejbs/CallSapAdapter
12/08/2014 09:26:06.472    Information    Application attempting to send an XI message asynchronously using connection HTTP_AAE_http://sap.com/xi/XI/System
12/08/2014 09:26:06.478    Information    Trying to put the message into the send queue
12/08/2014 09:26:06.499    Information    Message successfully put into the queue
12/08/2014 09:26:06.499    Information    The message was successfully retrieved from the send queue
12/08/2014 09:26:06.499    Information    The application sent the message asynchronously using connection HTTP_AAE_http://sap.com/xi/XI/System. Returning to application
12/08/2014 09:26:06.500    Information    HTTP adapter successfully processed interface PlantResponse_Out
12/08/2014 09:26:06.501    Information    Trying to put the message into the send queue
12/08/2014 09:26:06.501    Information    Message status set to DLNG
12/08/2014 09:26:06.513    Information    Message successfully put into the queue
12/08/2014 09:26:06.513    Information    The message was successfully retrieved from the send queue
12/08/2014 09:26:06.515    Information    Message status set to DLNG
12/08/2014 09:26:06.533    Information    Trying to put the message into the send queue
12/08/2014 09:26:06.548    Information    Message successfully put into the queue
12/08/2014 09:26:06.548    Information    The message was successfully retrieved from the send queue
12/08/2014 09:26:06.550    Information    Message status set to DLNG
12/08/2014 09:26:06.551    Information    Executing Request Mapping "http://styrolution.com/pi/Integration/PlantMaintenance/PlantResponse/PlantResponse_to_Zcol_Basf_Inotif02_Inotif01_Zeupm_Inotif01_Ludwigshafen" (SWCV 085d2320fb3b11e2cc0bf4d50aea8c33)
12/08/2014 09:26:06.563    Information    Trying to put the message into the send queue
12/08/2014 09:26:06.571    Information    Message successfully put into the queue
12/08/2014 09:26:06.571    Information    The message was successfully retrieved from the send queue
12/08/2014 09:26:06.573    Information    Trying to put the message into the send queue
12/08/2014 09:26:06.573    Information    Message status set to DLNG
12/08/2014 09:26:06.580    Information    Message successfully put into the queue
12/08/2014 09:26:06.580    Information    The message was successfully retrieved from the send queue
12/08/2014 09:26:06.582    Information    Message status set to DLNG
12/08/2014 09:26:06.583    Information    Delivering to channel: IDOC_R
12/08/2014 09:26:06.583    Information    MP: processing local module localejbs/IDocOutboundModuleBean
12/08/2014 09:26:06.583    Information    XI message received for processing
12/08/2014 09:26:06.587    Information    Sender Party: Sender Service:SXX Receiver Party: Receiver Service:RXXCLNT200 Communication Channel:IDOC_R
12/08/2014 09:26:06.588    Information    IDOC metadata repository is RXX
12/08/2014 09:26:06.589    Information    Control record is not mandatory - control record will be added if not available
12/08/2014 09:26:06.590    Information    XML Payload parsed to idoc document list with number of idocs: 1
12/08/2014 09:26:06.595    Information    Sending idoc message to receiver R/3 system with TID XIh{lsRI824UITC0009OalmW
12/08/2014 09:26:06.813    Information    IDOC message sent to receiver R/3 system
12/08/2014 09:26:06.817    Information    TID XIh{lsRI824UITC0009OalmW confirmed
12/08/2014 09:26:06.819    Information    Message was successfully transmitted to endpoint <local> using connection HTTP_AAE_http://sap.com/xi/XI/System
12/08/2014 09:26:06.821    Information    Message status set to DLVD
Thanks & regards
Dingjun

Hello All,
Thanks for the prompt responses and with best information, But apart from this,I need how an adapter(File/JDBC) works along
with AAE in PI 7.1. When does the job of File/JDBC adapter is completed and resumed in a synchronous case. That is,
File Location --> file adapter -
> XI(AAE) --> JDBC adapter --> database.
In file sender, lets suppose we have synchronous settings(Best Effort). Now when File adapter's job is finished and when it
gets back the response back from XI. Please explain step by step from starting(file adapter polling the location) to getting the response coming back to XI.
Regards,
Achari.

Progress bars and threads...

Hey...
I've had many replies on my previous questions about progress bars. I thank everybody who was concerned. But i still don't know if i have to use a seperate thread to run the progress of my bar... Can i have a progress bar in a Frame updating the progress of some file loading without having it in a seperate thread ?? I mean if i have an application that runs a Frame, this frame has buttons, textfields and other stuff, one of the buttons uploads a file chosen to a server. Now i also have a progress bar, can it keep progress of the upload without having it in a seperate thread ?? If yes, then how. Cause i can't seem to make it move unless the upload is done... I also want to know that, if i have to use a seperate thread, do i have to use something called a SwingWorker ???

Here's the deal....
If you don't use a separate thread, then your file upload process will block the main ("dispatch") thread until it has finished. This means that the GUI will not be updated to show your ever-growing progress bar. You'll see the bar start at zero, and then nothing will happen until the file finishes uploading, and then the bar will show 100%.
Even if you are updating the bar in a loop, changes will still not be displayed, because repaint requests get put on a queue, and they are executed only when the VM decides it has time to do them (ie when it's finished doing your file upload!)
So, yes you do need to use a thread to load your file and change your progress bar. and if that thread will update any GUI elements directly, start it by calling SwingUtilities.invokeLater(myThread). Also, make sure you do "myThread.setPriority(3);" (or thereabouts - make sure it's less than 5) on the thread before invoking it. This way, the main thread that updates the gui will have priority, so will update regularly.

JMS Adapter , EDI and SEEBURGER Adapters

HI All
Please briefly explain JMS adapter ,EDI and seeburger adapter .could you please the correlated links

Hello
This thread has been locked by a Moderator. Please familiarise yourself with the forum Rules of Engagement before posting questions.
Rules of Engagement
https://wiki.sdn.sap.com/wiki/display/HOME/RulesofEngagement
Regards
Mark Smyth
XI/PI Moderator

Set up MQ interconnect log queue and MQ sequence/transaction id queues?

Does the log queue and sequence queue get set up in standard MQ server install?
Can the log and sequence queue also point to the actual queue?
It looks like a standard MQ client only needs the channel and send queue (the actual queue that is the destination for inbound (to the MQ server) messages).
Whereas the oracle mq adapter requires 3 inbound queues; actual, log, and transaction id.
Any help/more documentation/additional install guides/tips would be greatly appreciated.
thanks

Yes you do. The log queue must be created on the MQ side and you specify it in your link set tup. Oracle requires that to guarantee "deliver once and only once" in case of errors. If you don't specify that, Oracle will warn you. It's easy for the MQ admin to create. Just ask them.

Difference between Adaptive RFC and Adaptive RFC2 model

Hi,
What is the difference between Adaptive RFC and Adaptive RFC2 model ?
Regards,
Sunaina Reddy T

Hi
Main difference are
1.JCo 3.0 to connect to SAP Systems
2.Improved connection management
3.Faster performance
4.Lower memory consumption
5.SystemLandscape API is no longer needed
6.Compatible to ARFC1 to allow easy migration
Check the thread for further input
1.[Whatu2019s New in Web Dynpro Java?|https://www.sdn.sap.com/irj/scn/go/portal/prtroot/docs/library/uuid/8062c7d3-c86f-2b10-4894-a9a323da20b3]
Best Regards
Satish Kumar

TestStand Execution and Thread Limits?

Is there a limit to the number of executions and threads that TestStand can support? If not, what is the "practical" limit of Test Sockets for a modern PC running 64-bit Windows? I understand this is a pretty broad question that is probably based upon the contents of the sequences along with the PC resources (RAM, CPU). However, I am curious to find out what would be suggested/experienced.
I am looking at implementing "monitoring" ten to hundreds of signals on a system in parallel. The loop rate for each signal read is specified in a configuration file that can change at runtime. To implement this in TestStand, a sequence in a new thread would be required for each unique loop rate (signal monitor). This would possibly create hundreds of threads under an execution.
Any advice would be appreciated!
Thanks
CLA, CTA

There will be a limit... probably when you run out of free memory blocks of a sufficient size. This will vary according to what modules are loaded, what has been allocated and freed at what locations, whether large address awareness is enabled, etc. Your best bet is to test how many TestStand threads you can create in your system within an application that is as close as possible to your desired application and then estimate the reliably obtainable maximum number threads as a fraction of that.

C2960: ingress priority-queue and CoS/DSCP mapping

I have read the C2960 manuals for a few times, but I don't seem to find an answer.
You can enable the priority queue on the ingress interface which has two available input queues for QoS-services.
What seems to be missing from the manuals is that how can I map/direct the traffic (CoS & DSCP values) to the priority queue?
I can map traffic to two other WRR (SRR) queues with command: "mls qos srr-queue input cos-map queue queue-id threshold threshold-id cos1...cos8", but there is nothing about the priority queue.
So is the CoS/DSCP values for the priority queue fixed or how/what traffic is directed to it?
Thanks,
Miska

What seems to be missing from the manuals is that how can I map/direct the traffic (CoS & DSCP values) to the priority queue?
A: By default, the priority queue is queue 2, and 10 percent of the bandwidth is allocated to it.
?For queue-id, the range is 1 to 2.
Map DSCP or CoS values to an ingress queue and to a threshold ID.
By default, DSCP values 0-39 and 48-63 are mapped to queue 1 and threshold 1. DSCP values 40-47 are mapped to queue 2 and threshold 1.
By default, CoS values 0-4, 6, and 7 are mapped to queue 1 and threshold 1. CoS value 5 is mapped to queue 2 and threshold 1.
?For queue-id, the range is 1 to 2.
?For threshold-id, the range is 1 to 3. The drop-threshold percentage for threshold 3 is predefined. It is set to the queue-full state.
?For dscp1...dscp8, enter up to eight values, and separate each value with a space. The range is 0 to 63.
?For cos1...cos8, enter up to eight values, and separate each value with a space. The range is 0 to 7.
What above is saying is queue 2 is priority queue, you can use the CoS/DSCP mapping to queue to move certain CoS?DSCP to either queue 1 or queue 2.
I can map traffic to two other WRR (SRR) queues with command: "mls qos srr-queue input cos-map queue queue-id threshold threshold-id cos1...cos8", but there is nothing about the priority queue.
So is the CoS/DSCP values for the priority queue fixed or how/what traffic is directed to it?
A; See above.
For more info:
http://www.cisco.com/univercd/cc/td/doc/product/lan/cat2960/12225see/scg/swqos.htm#wp1161845
Please rate all posts.

MDBs and threads

We are using Weblogic SP2 with MQ Series and MDBs. We have MDBs reading the queues
and calling Clarify application APIs (Clarify Business Objects). Everything works
fine on low volumes, but as the number of messages increase, weblogic stops responding
and the beans hang. There is no error or exception!
We have MDBS configured at 1min and 8max. Also, Java APIs have synchronized methods
to provide thread-safe access.
Any ideas?

Start by taking periodic thread thumbs and go from there... Most likely
cause is deadlock:(
"Prashant" <[email protected]> wrote in message
news:[email protected]..
>
We are using Weblogic SP2 with MQ Series and MDBs. We have MDBs readingthe queues
and calling Clarify application APIs (Clarify Business Objects).Everything works
fine on low volumes, but as the number of messages increase, weblogicstops responding
and the beans hang. There is no error or exception!
We have MDBS configured at 1min and 8max. Also, Java APIs havesynchronized methods
to provide thread-safe access.
Any ideas?

How to create JMS Queues and Topics - OAS 10.1.3?

Hi,
Is there a way to create JMS Queues and Topics via the command line for OAS 10.1.3, say through admin-client.jar ?
I have read the "Managing Adminitered Objects" chapter from the "Oracle Containers for J2EE Resource Adapter" book and could not find the information I am looking for.
Thanks.
-Sachin

Hi Sachin --
In the 10.1.3.1 release, a set of new commands have been added to admin_client.jar to support the command line creation of JMS resources:
-addJMSConnectionFactory -domain <domain> -jndiLocation <jndiLocation> [options]
-removeJMSConnectionFactory -jndiLocation <jndiLocation>
-getJMSConnectionFactories
-addDestination -domain <domain> -name <name> -jndiLocation <jndiLocation> [options]
-removeDestination -name <name> [options]
-getDestinations
You can read about it in the doc:
http://download-west.oracle.com/docs/cd/B31017_01/web.1013/b28950/adminclient.htm#CHDDFJJA
cheers
-steve-

Adapter queues and thread allocation

Similar Messages

Maybe you are looking for