Assigning MBeans to own execute queue
We are trying to monitor the internal activities of the WLS by using the
MBeans (JMX).
However, when load is high we see that the MBean access halts or is
seriously impared. Ideally we would like the MBeans (or the classes that
access them) to run in their own execute queue.
We access MBeans from stand-alone Java classes that then use JNDI to locate
the MBeanHome and make calls from there.
Is there any way to get this design to run in its own execute queue?
If not, am I right in thinking if I put a session (facade) bean in the
design I could designate that this bean ran in its own queue?
TIA
Ed
Dimitri I. Rakitine wrote:
That should work - if I understand correctly how this works, request is placed on an
appropriate queue when it first arrives.
I have a question - it looks like console servlets are configured to run on
'__weblogic_admin_html_queue', but what is configured to execute on
'__weblogic_admin_rmi_queue' ?
Managed servers could make RMI calls to admin server to get config
MBeans? etc....
In weblogic.developer.interest.performance Ed Barrett <[email protected]> wrote:
We are trying to monitor the internal activities of the WLS by using the
MBeans (JMX).
However, when load is high we see that the MBean access halts or is
seriously impared. Ideally we would like the MBeans (or the classes that
access them) to run in their own execute queue.
We access MBeans from stand-alone Java classes that then use JNDI to locate
the MBeanHome and make calls from there.
Is there any way to get this design to run in its own execute queue?
If not, am I right in thinking if I put a session (facade) bean in the
design I could designate that this bean ran in its own queue?
TIA
Ed
Similar Messages
-
Problem while assigning Servlet to Execute queue
Hi,
We implement a IMS S-CSCF on BEA container 2.2.
A call for us may be several times reentrant. Meaning that for one call we have to handle at least three different dialogs but in major use cases, it will be 6 or 7 for our next deployment.
We are currently facing to bottlenecks problems and BEA support advised me to try to check if assigning servelt to execute queue could improve the preformances.
I just wanted to make a try with one of our servlet but I have an exception java.lang.NoClassDefFoundError: org/jaxen/JaxenException.
Here is what I did:
In web.xml I have:
<web-app>
<distributable/>
<servlet>
<servlet-name>register</servlet-name>
<servlet-class>com.netcentrex.scscf.registrar.RegistrarServlet</servlet-class>
<init-param>
<param-name>wl-dispatch-policy</param-name>
<param-value>REGISTRARQueue</param-value>
</init-param>
<load-on-startup/>
</servlet>
<listener>
<listener-class>com.bea.wcp.sip.engine.server.ContextEventHook</listener-class>
</listener>
</web-app>
In sip.xml I have:
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE sip-app
PUBLIC "-//Java Community Process//DTD SIP Application 1.0//EN"
"http://www.jcp.org/dtd/sip-app_1_0.dtd">
<sip-app>
<display-name>SIP Servlet based Registrar</display-name>
<distributable/>
<context-param>
<param-name>usc.configFile</param-name>
<param-value>/etc/neotip/usc.ini</param-value>
</context-param>
<context-param>
<param-name>icscf.configFile</param-name>
<param-value>/etc/neotip/icscf.ini</param-value>
</context-param>
<context-param>
<param-name>usc.dnsSrvFile</param-name>
<param-value>/etc/neotip/uscDnsSrv.cache</param-value>
</context-param>
<listener>
<listener-class>com.netcentrex.scscf.session.SessionServlet</listener-class>
</listener>
<servlet>
<servlet-name>register</servlet-name>
<servlet-class>com.netcentrex.scscf.registrar.RegistrarServlet</servlet-class>
<load-on-startup/>
</servlet>
<servlet>
<servlet-name>session</servlet-name>
<servlet-class>com.netcentrex.scscf.session.SessionServlet</servlet-class>
<load-on-startup/>
</servlet>
<servlet>
<servlet-name>route</servlet-name>
<servlet-class>com.netcentrex.scscf.fork.RouteServlet</servlet-class>
<load-on-startup/>
</servlet>
<servlet>
<servlet-name>icscf</servlet-name>
<servlet-class>com.netcentrex.scscf.embeddedI.IEmbeddedServlet</servlet-class>
<load-on-startup/>
</servlet>
<servlet-mapping>
<servlet-name>register</servlet-name>
<pattern>
<equal>
<var>request.method</var>
<value>REGISTER</value>
</equal>
</pattern>
</servlet-mapping>
<servlet-mapping>
<servlet-name>session</servlet-name>
<pattern>
<and>
<not>
<equal>
<var>request.method</var>
<value>REGISTER</value>
</equal>
</not>
<not>
<exists>
<var>request.uri.param.uscroute</var>
</exists>
</not>
<not>
<exists>
<var>request.uri.param.icscfroute</var>
</exists>
</not>
</and>
</pattern>
</servlet-mapping>
<servlet-mapping>
<servlet-name>route</servlet-name>
<pattern>
<and>
<not>
<equal>
<var>request.method</var>
<value>REGISTER</value>
</equal>
</not>
<exists>
<var>request.uri.param.uscroute</var>
</exists>
</and>
</pattern>
</servlet-mapping>
<servlet-mapping>
<servlet-name>icscf</servlet-name>
<pattern>
<and>
<not>
<equal>
<var>request.method</var>
<value>REGISTER</value>
</equal>
</not>
<exists>
<var>request.uri.param.icscfroute</var>
</exists>
</and>
</pattern>
</servlet-mapping>
<session-config>
<session-timeout>30</session-timeout>
</session-config>
</sip-app>
In msi-config.xml, we have:
<Server Cluster="BEA_ENGINE_TIER_CLUST" EnabledForDomainLog="false"
ListenAddress="192.168.150.26" ListenPort="8001" Name="engine1"
NativeIOEnabled="true" NetworkAccessPoints="sipchannel"
ServerVersion="8.1.5.0" StdoutDebugEnabled="true"
StdoutEnabled="true" StdoutSeverityLevel="64">
<NetworkAccessPoint ListenAddress="192.168.150.34"
ListenPort="5060" Name="sipchannel" Protocol="sip"
PublicAddress="192.168.150.34" SourcePorts="41000-42000"/>
<SSL Enabled="false" IdentityAndTrustLocations="KeyStores" Name="engine1"/>
<Log FileCount="10" FileMinSize="60000" Name="engine1" NumberOfFilesLimited="true"/>
<ExecuteQueue Name="sip.transport.Default" QueueLength="2048" ThreadCount="25"/>
<ExecuteQueue Name="REGISTRARQueue" QueueLength="2048" ThreadCount="25"/>
<ExecuteQueue Name="sip.transport.Default" QueueLength="2048" ThreadCount="25"/>
<ExecuteQueue Name="weblogic.kernel.Default"/>
<ExecuteQueue Name="sip.timer.Default" QueueLength="2048" ThreadCount="25"/>
<ExecuteQueue Name="sip.tracing.local" QueueLength="1024"
ThreadCount="1" ThreadsMaximum="1" ThreadsMinimum="1"/>
<ExecuteQueue Name="sip.tracing.domain" QueueLength="1024"
ThreadCount="1" ThreadsMaximum="1" ThreadsMinimum="1"/>
</Server>
Then when I launch a SIP REGISTRATION operation I get the following trace:
2007-05-30 17:37:03,649 ERROR RegistrarServlet [ExecuteThread: '23' for queue: 'sip.transport.Default'] : Error while processing request: REGISTER sip:netcentrexbench.net SIP/2.0
P-Charging-Vector: icid-value=AyretyU0dm+1970294253=0465138796
P-Access-Network-Info: ADSL;dsl-location="netcentrexbench.net;1.2.3.4"
CSeq: 2 REGISTER
Call-ID: [email protected]
Supported: path
Via: SIP/2.0/UDP 192.168.254.56:5063;branch=z9hG4bK-19606-1-2
From: dench0 <sip:[email protected]>;tag=1
Allow: INVITE, ACK, CANCEL, OPTIONS, BYE, REFER, NOTIFY, MESSAGE, SUBSCRIBE, INFO
To: dench0 <sip:[email protected]>
Authorization: Digest username="dench0",realm="netcentrexbench.net",cnonce="6b8b4567",nc=00000001,qop=auth,uri="sip:192.168.150.34:5060",nonce="NQei3RIBAAA=le+kkCcrZcahXxQ7M6r7dw==",response="b824a93822b267ddfaa07f4039d99d95",algorithm=MD5
Content-Length: 0
Contact: sip:[email protected]:5063
User-Agent: LYNX_BENCH
Path: <sip:[email protected]:5063;lr>
Require: path
Expires: 36000
Max-Forwards: 70
java.lang.NoClassDefFoundError: org/jaxen/JaxenException
at org.dom4j.DocumentFactory.createXPath(DocumentFactory.java:230)
My first doubt is that I think I should get "ERROR RegistrarServlet [ExecuteThread: '23' for queue: 'REGISTRARQueue'] " instead of "ERROR RegistrarServlet [ExecuteThread: '23' for queue: 'sip.transport.Default'] "
First question: did I miss something or did something wrong in the configuration?
Second question: is it really possible to assign a single servlet (from an application containing many other servlets) to a dediacted execute queue?
Third question: any help is welcome :-)
Thanks in advance,
Lionel,At teh end I found the basic :-) problem on my exception but anyway , the used execute queue for REGISTER is always sip.transport.default
Help on this point is welcome!
Thanks -
Problem in assigning the execute queue to EJB.
I have a enterprise application in that a have many web application.I want to assign
the low priority threads to one web application e.g. "A". The "A" is using one
ejb to execute the some task.
I am making a execute queue with name let "TestQueue"
and assigning it to my servlet thru these entries in
weblogic-ejb-jar :-
<weblogic-enterprise-bean>
<ejb-name>InferenceEngine</ejb-name>
<reference-descriptor>
<resource-description>
<res-ref-name>jdbc/MyDB</res-ref-name>
<jndi-name>jdbc/MyDB</jndi-name>
</resource-description>
</reference-descriptor>
<jndi-name>ejb/InferenceEngine</jndi-name>
<dispatch-policy>TestQueue</dispatch-policy>
</weblogic-enterprise-bean>
Here is my code of config.xml file for execute queue.
<?xml version="1.0" encoding="UTF-8"?>
<Domain ConfigurationVersion="8.1.0.0" Name="ACM_1">
<Server ListenAddress="" ListenPort="8003" Name="ACM"
NativeIOEnabled="true" ReliableDeliveryPolicy="RMDefaultPolicy" ServerVersion="8.1.1.0">
<SSL Enabled="false" HostnameVerificationIgnored="false"
IdentityAndTrustLocations="KeyStores" Name="ACM"/>
<ExecuteQueue Name="TestQueue" ThreadPriority="1"/>
</Server>
I tried the ejbc command line argument -dispatchPolicy TestQueue also. But it
is also not sending request to my TestQueue.One thing to note is the execute queue is chosen when the request enters
the server. So if a webapp is assigned to Queue-1 and an ejb to
Queue-2, and the webapp calls the ejb, the request will remain on queue-1.
-- Rob
Sunil Kumar Mehta wrote:
I have a enterprise application in that a have many web application.I want to assign
the low priority threads to one web application e.g. "A". The "A" is using one
ejb to execute the some task.
I am making a execute queue with name let "TestQueue"
and assigning it to my servlet thru these entries in
weblogic-ejb-jar :-
<weblogic-enterprise-bean>
<ejb-name>InferenceEngine</ejb-name>
<reference-descriptor>
<resource-description>
<res-ref-name>jdbc/MyDB</res-ref-name>
<jndi-name>jdbc/MyDB</jndi-name>
</resource-description>
</reference-descriptor>
<jndi-name>ejb/InferenceEngine</jndi-name>
<dispatch-policy>TestQueue</dispatch-policy>
</weblogic-enterprise-bean>
Here is my code of config.xml file for execute queue.
<?xml version="1.0" encoding="UTF-8"?>
<Domain ConfigurationVersion="8.1.0.0" Name="ACM_1">
<Server ListenAddress="" ListenPort="8003" Name="ACM"
NativeIOEnabled="true" ReliableDeliveryPolicy="RMDefaultPolicy" ServerVersion="8.1.1.0">
<SSL Enabled="false" HostnameVerificationIgnored="false"
IdentityAndTrustLocations="KeyStores" Name="ACM"/>
<ExecuteQueue Name="TestQueue" ThreadPriority="1"/>
</Server>
I tried the ejbc command line argument -dispatchPolicy TestQueue also. But it
is also not sending request to my TestQueue. -
Different execute queues with different thread priorities
We are running WebLogic 6.1, SP2 on Solaris 2.8 and I have a question related to the
use of Execute queues
I'm aware that the administration console servlets, and the rest of requests given
to the server run in separate execution queues each with their own pool of threads.
There is an <executequeue> element in the config.xml that allows you to change the
parameters of the default execute queue that most requests are handled in.
My question is is it possible to create new execute queues, each with a given number
of threads running at a different priority. Then one could assign a particular WAR
deployment to a queue and hence control the priority that certain requests run at
governed by the deployment that received the initial request.
Essentially the problem that I'm trying to address is that I have a JAR with MDBs
in it. I also have JARs with Entity and Session EJBs and WARs. They are all using
the same pool of threads to execute. I want to be able to throttle the MDBs as they
receive messages so they don't take over the machine and interfere with interactive
use. Now I can limit the number of MDBs in the MDB pool, but then if the machine
is not being used by users, then the machine resources are wasted only allowing one
MDB to be active at a time when there are a lot of messages in the queue.
What I want to be able to do is to have as many MDBs run as I can, up to a certain
limit, say 15 for example, but not to pull from the same thread pool/execute queue
as the requests for URLs in the WAR. Also I want the MDB threads to run at a much
lower priority and always give preference to interactive requests.
So interactive threads will always win and grab the processors if interactive use
is high, but if it's not, the MDBs can run instead.
Regs,
PaulHi,
I had a chat with BEA support about my issue. You can create multiple ExecuteQueues
and assign them to servlet definitions in WARs in the web.xml, and with an option
to the EJBC compiler, you can assign a particular class of EJB to an ExecuteQeueue.
For details see http://e-docs.bea.com/wls/docs61/perform/AppTuning.html#1106284
Presumably it's the thread of the initial request that counts, i.e. if a WAR calls
an EJB, then it's the thread in the execute queue of the WAR that will be used right
the way through the request, not that of the EJB being called, assuming they are
both deployed in the same WebLogic instance. If they were in different instances,
you would obviously have a thread from the WAR queue on one instance, and a thread
from the EJB queue on another instance.
If a plain Java command-line client connected directly to an EJB, then it would be
the thread from the ExecuteQueue of the EJB that would get used.
I've been told that MDBs always run in the default queue, so it seems that the way
for me to do what I want is just to create an ExecuteQueue, set the priority of the
threads in this queue to be higher than that of the default queue, and assign this
queue to my WARs. I can leave my EJBs running in the default queue.
Regs,
Paul
"Dimitri I. Rakitine" <[email protected]> wrote:
Don't know about 6.1 (One can find out very easily what the possible config
parameters
are by deploting
http://dima.dhs.org/misc/listMBeans.jsp and
http://dima.dhs.org/misc/showMBean.jsp
abd looking at the XXXConfig mbeans), but in 7.0 there are quite a few
things which
can be tuned in the executequeue config (and priority is one of them. also
interesting
is that it is possible to configure it to increase number of execute threads
when queue
length exceeds some specified threshold) :
QueueLength
Returns the maximum length of this queue.
ThreadPriority
Returns the priority of the threads associated with this queue.
ThreadCount
Returns the number of threads assigned to this queue.
QueueLengthThresholdPercent
Returns the threshold percent for length of this queue, set in QueueLength.
ThreadsIncrease
Returns the number of threads to grow when a queue is within QueueLengthThresholdPercent
of the set QueueLength.
ThreadsMaximum
Returns the maximum number of threads in the pool.
ThreadsMinimum
Returns the minimum number of threads in the pool.
Dimitri -
How can I tell how long a request has been sitting on an execute queue?
Is it possible to determine (in WL 8.1 and/or 10.3) how long an HTTP request has been sitting on the execute queue before being assigned to a servlet?
Under heavy load, I'm finding that the caller times out by the time the servlet tries to write the response because the request has been waiting on the execute queue for too long. To prevent doing all the work to process the request, I want to be able to see how long the request has been waiting to determine whether it's worth processing the request.
Currently I use the getExecuteThreadTotalCount() method in ExecuteQueueRuntimeMBean and assume that if the count is high, the request has probably been waiting for a while, but getting the actual time on the queue would be much better.
Thanks.Further investigation and a bunch of speed tests later, I'm even more puzzled. Between 7 pm and 9pm this evening, download speed has been consistently at 112kb, except for the 1 in 10 occasions where it hits 1.5mb, for 30 or 40 seconds, before dropping back to 112kb.
Does anyone have any ideas what is going on?
thanks in advance
Ben -
How to find the Current threads idle in each execute queue
Hi,
Currently I am able to get the idle threads in the default queue.
I would like to find out if it is possible to find out the no of threads
that are idle in each of the execute queues in the wls server.
Thanks
SenthilSenthil Kumar wrote:
Hi,
Currently I am able to get the idle threads in the default queue.
I would like to find out if it is possible to find out the no of threads
that are idle in each of the execute queues in the wls server.
Thanks
Senthil
Senthil, the easiest way from the command-line is by using a
ExecuteQueueRutime MBean. You could use the simple syntax below or use
the more specific domain oject name. You could also omit the -property
option and get a list of every attribute for the bean.
%java weblogic.Admin -url t3://127.0.0.1:7001 -username system -password
weblogic -get -pretty -type ExecuteQueueRuntime -property
ExecuteThreadCurrentIdleCount
HTH
~Ryan -
Hi,
We are working with weblogic 8.1. We have configured one Execute Queue for our
EJB, and modified deployment desriptor weblogic-ejb-jar.xml using this queue.
In our application we are using scheduler program(Quartz) to invoke this EJB,
but the calls are not going in to this Execute queue, instead it goes to the default
queue.
But when i call EJB method from other client(different VM), calls goes through
the execute queue.
For your information, we haven't modified <enable-call-by-reference> tag, which
means this option is set to default value false.
Please let me know if i am missing something?
Regards
NareshHi,
To assign an EJB object to a configured execute queue, use the new dispatch-policy element in weblogic-ejb-jar.xml.
Ex : dispatch-policy>queue_name</dispatch-policy>
If no dispatch-policy is specified, or the specified dispatch-policy refers to a nonexistent server execute thread pool, then the server's default execute thread pool is used instead.
Go through the following links and find the information on this issue.
http://e-docs.bea.com/wls/docs81/perform/AppTuning.html#1104273
http://e-docs.bea.com/wls/docs81/ejb/DDreference-ejb-jar.html#dispatch-policy
Regards
Anilkumar kari -
Creating Execute Queue through WLConfig ANT tasks
We are planning to create a new Execute Queue for a existing server in weblogic 8.1. When we create a new execute queue through the web admin console it add the following line within the Server tag in config.xml.
<?xml version="1.0" encoding="UTF-8"?>
<Domain Name="weblogicdomain">
<Server ListenAddress="" ListenPort="6110" Name="servername">
<ExecuteQueue Name="FileDownloadsExecuteQueue" ThreadCount="1"/>
</Server>
....Other resources....
</Domain>
Now we have to create this execute queue NOT through the web admin console but through wlconfig ANT tasks. How should we create this? There are samples ANT tasks to add a resource (JDBC connection pool etc) directly within the Domain tag of the config.xml but our requirement is to add a new element (ExecuteQueue) within the existing server tag. As the server is already created we cannot create server mbean and nest it with one more create of execute queue. We have to attach the new execute queue to the current existing server.
Can we set a new element (not a attribute) to the server by quering or any other method?
Thanks.I haven't tried this with wlconfig, but I was able to accomplish it by invoking weblogic.Admin from the <java> task:
<!-- Create the ExecuteQueue -->
<java classname="weblogic.Admin" fork="true">
<arg line="-url ${deploy.server.name}:${deploy.server.port}"/>
<arg line="-username ${deploy.username}"/>
<arg line="-password ${deploy.password}"/>
<arg value="CREATE"/>
<arg line="-mbean portal:Type=ExecuteQueue,Name=newRenderQueue,Server=${deploy.target}"/>
<classpath>
<pathelement location="${weblogic.home.dir}/server/lib/weblogic.jar"/>
</classpath>
</java>
<!-- Set the ThreadCount -->
<java classname="weblogic.Admin" fork="true">
<arg line="-url ${deploy.server.name}:${deploy.server.port}"/>
<arg line="-username ${deploy.username}"/>
<arg line="-password ${deploy.password}"/>
<arg value="SET"/>
<arg line="-mbean portal:Type=ExecuteQueue,Name=newRenderQueue,Server=${deploy.target}"/>
<arg line="-property ThreadCount 5"/>
<classpath>
<pathelement location="${weblogic.home.dir}/server/lib/weblogic.jar"/>
</classpath>
</java>
Hope that helps,
George -
Execute Queue configguration reset.
Suppose if one has a WL domain with admin server and managed server as01 running
on host1, managed server as02 on host2, as03 on host3 and as04 on host4.
While rstarting the domain after a code update...
1. started admin server on host1
2. started as01 on host1
3. started as02 on host2
4. started as03 on host3
5. accidently tried starting as04 on host1 as opposed to host4.
Wl server bailed out after attempting to start as04 on host1 with the following
exception:
<Nov 12, 2003 8:49:57 AM PST> <Emergency> <WebLogicServer> <000350> <Unable to
create a server socket on Channel Default for:
113.130.234.101, port: 8001. java.net.BindException: Cannot assign requested address
Perhaps the address 113.130.234.101 is in
correct or another process is using port 8001.>
<Nov 12, 2003 8:49:58 AM PST> <Emergency> <Security> <090085> <Server failed to
bind to any port such that it is not reachable
from the console.>
<Nov 12, 2003 8:49:58 AM PST> <Emergency> <WebLogicServer> <000342> <Unable to
initialize the server: Fatal initialization exc
eption>
The WebLogic Server did not start up properly.
Reason: Fatal initialization exception
Exception: weblogic.rmi.extensions.RemoteRuntimeException: Unexpected Exception
- with nested exception:
As a side affect, WL admin server reset execute queue the config setting for all
the managed servers. The config file got changed from
<ExecuteQueue Name="default" ThreadCount="300"
ThreadPriority="6" ThreadsIncrease="10"/>
<IIOP Name="as01"/>
to
<ExecuteQueue Name="default" ThreadCount="15"/>
in the config.xml file after the failed start of as04.
Any idea why WL did such a thing?Hi Dhilip,
Please post this question to the following newsgroup:
weblogic.developer.interest.general
Vitaly Yurik
"Dhilip Krishnan" <[email protected]> wrote:
>
Suppose if one has a WL domain with admin server and managed server as01
running
on host1, managed server as02 on host2, as03 on host3 and as04 on host4.
While rstarting the domain after a code update...
1. started admin server on host1
2. started as01 on host1
3. started as02 on host2
4. started as03 on host3
5. accidently tried starting as04 on host1 as opposed to host4.
Wl server bailed out after attempting to start as04 on host1 with the
following
exception:
<Nov 12, 2003 8:49:57 AM PST> <Emergency> <WebLogicServer> <000350> <Unable
to
create a server socket on Channel Default for:
113.130.234.101, port: 8001. java.net.BindException: Cannot assign requested
address
Perhaps the address 113.130.234.101 is in
correct or another process is using port 8001.>
<Nov 12, 2003 8:49:58 AM PST> <Emergency> <Security> <090085> <Server
failed to
bind to any port such that it is not reachable
from the console.>
<Nov 12, 2003 8:49:58 AM PST> <Emergency> <WebLogicServer> <000342> <Unable
to
initialize the server: Fatal initialization exc
eption>
The WebLogic Server did not start up properly.
Reason: Fatal initialization exception
Exception: weblogic.rmi.extensions.RemoteRuntimeException: Unexpected
Exception
- with nested exception:
As a side affect, WL admin server reset execute queue the config setting
for all
the managed servers. The config file got changed from
<ExecuteQueue Name="default" ThreadCount="300"
ThreadPriority="6" ThreadsIncrease="10"/>
<IIOP Name="as01"/>
to
<ExecuteQueue Name="default" ThreadCount="15"/>
in the config.xml file after the failed start of as04.
Any idea why WL did such a thing? -
What are the drawbacks of creating execute queues
Hi thread Gurus
What are the drawbacks of using execute queues. I understand that if queues are not used they stay idle but on solaris the SunOS scheduler will only give active
threads CPU time. what ar the other resources being wasted in the process.
I will appreciate if you can farward me to some documentation on this apart from performance docs on BEA site.
Regards
ParminderHi Parminder,
BEA docs cover tweaking exec queues pretty extensively.
Could you provide more details on what kind of information
you are looking for or which concerns do you have?
Regards,
Slava imeshev
"Parminder" <[email protected]> wrote in message
news:3ef86b40$[email protected]..
Hi thread Gurus
What are the drawbacks of using execute queues. I understand that ifqueues are not used they stay idle but on solaris the SunOS scheduler will
only give active
threads CPU time. what ar the other resources being wasted in the process.
I will appreciate if you can farward me to some documentation on thisapart from performance docs on BEA site.
Regards
Parminder -
Identifying Execute Queue Threads as Socket Readers
Hi All,
Preface: I am trying to identify at any given instance the percentage of
socket reader threads that are in use - this will help me to identify when
all threads are in use and unable to service user requests. Thus my goal is
to identify threads that are socket readers and whether they are active or
idle.
Through JMX I can obtain an instance of each
weblogic.management.runtime.ExecuteThread, which lists the same information
that you see in the console if you drill down to the thread level, for
example:
mydomain> Servers> myserver> Active Execute Queues> default> Execute Threads
That information includes:
- thread number
- total requests
- last request
- current request
- transaction
- user
- is the thread idle?
Now my question is from this information can you identify which of these
threads are being used as socket readers?
Here are some excerpts from my JMX calls:
Name: ExecuteThread: '0' for queue: 'default'
Execute Queue Runtime Name: default
Current Request: null
Last Request: Http Request: /bookstore/en/authors/showauthors.jsp
User: null
Is Idle?: true
Name: ExecuteThread: '15' for queue: 'default'
Execute Queue Runtime Name:default
Current Request: null
Last Request: Scheduled Trigger
User: null
Is Idle?: true
Name: ExecuteThread: '26' for queue: 'default'
Execute Queue Runtime Name:default
Current Request: null
Last Request: ListenRequest for a new connection on: 'Socket
addr=127.0.0.1,port=2061,localport=7001]'
User: null
Is Idle?: true
Name: ExecuteThread: '59' for queue: 'default'
Execute Queue Runtime Name:default
Current Request: Socket Reader Request
Last Request:
weblogic.transaction.internal.ServerCoordinatorDescriptor$2@138786
User: null
Is Idle?: false
It would appear that the "ListenRequest" last request would identify a
socket reader, only I have 60 threads in my default execute queue with 70%
dedicated to socket readers and the number of threads that say their last
request was "ListenRequest ..." is only 5.. Furthermore there is only one
"Socket Reader Request" identified thread.
Are threads identified as socket readers permanently or constantly reused
for whatever purpose is required (with a cap of use based off of the socket
reader perspective)?
Thanks in advance for your help!
SteveHi Achhi,
Socket Reader Threads are the Threads which are responsible for reading the incoming request data. We can divide the Socket reader threads in two categories:
Pure Java Socket Reqders: The pure-Java socket reader implementation, where the socket reader threads continually poll all opened sockets to determine whether they contain data to be read, even if the sockets have no data to read. (From Performance Point This is Not Good that the Sockets will be in Opened Mode even if there is no Data to read)
Native Socket Readers: (These are the Dafault Socket Readers) The native IP socket reader provided by the host machine's operating system, where the socket reader threads target only active sockets that contain data to be read. Native socket readers never poll sockets because they are immediately notified when a socket needs to be serviced. For Better Performance You should always prefer using Native Socket Readers.
To Enable Native Socket Readers You can Login to AdminConsole--->Servers ---><SERVER_NAME>--->configuration (Tab)--->Tuning (SubTab)---> "Enable Native IO" (This checkbox must be checked)
Still If you want to use the Java Socket Readers .... still you can improve the performance of socket communication by configuring the proper number of socket reader threads for each server instance. For best performance, the number of socket reader threads in WebLogic Server should equal the potential maximum number of opened sockets.
Thanks
Jay SenSharma
http://weblogic-wonders.com/weblogic (WebLogic Wonders Are here) -
Execute queue thread count set back?
I have a script which automatically sets up execute queues (and their thread counts) for either an admin server or two managed servers in a cluster (depending on whether it's a standalone deployment). Upon building a domain, the script connects to a temporary server and sets everything up, at this point it appears all the queues and thread counts are set correctly. However once the build has completed and the generated config.xml is checked, three queues don't have any thread counts associated with them at all; these queues are:
ThreadQueue (Count)
wli.internal.ProcessInstanceInfo (15)
wli.internal.ProcessTracking (15)
wli.process.event.thread.pool (15)
The other queues have all their thread counts set correctly. Does anybody know how these queues are used and if this will have an impact on the above problem?No worries about this post, the reason has been solved.
Reason
The default thread count for execute queues in development mode is 15 threads, therefore if a queue is specified with this value it won't be set in the config.xml (when an execute queue doesn't have a ThreadCount attribute, the default is used instead). The trouble is that when the server is started we enter production mode where the default thread count is 25 threads instead. -
Specifying multiple JSP's in Execute Queue
Hi
Bea's documentation says that you can specify the execute queue for a
perticular JSP as follows:
<servlet>
<servlet-name>MainServlet</servlet-name>
<jsp-file>/myapplication/critical.jsp</jsp-file>
<init-param>
<param-name>wl-dispatch-policy</param-name>
<param-value>CriticalAppQueue</param-value>
</init-param>
</servlet>
Am wondering if there is a way to specify the execute queue for a
whole set of JSP's. Was trying wildcards in the jsp-file tag but
havent yet had success. I would like to tell weblogic to use the queue
I specify for all jsp's in a certain directory and I dont want to have
to create <servlet> tags for each one since there are many!.
If anyone has solved this or has any suggestions, please let me know.
ThanksDid you try in SM58( in the menu edit / Execute LUW's ? There send you to screen " Execute Calls Not Yet Executed" where you can select some parameters as excution date, destination user name..... and other options with check boxes.
I have used it and is good -
Any way to create Execute Queues or set Load Order using ANT?
Hello All,
Does anyone know whether wlconfig or some other Ant tasks can be used
in order to automate the creation of specific Execute Queues? We'd like
to script and package this so as to avoid having the user configure
queues through the admin console.
I'd also like to find out if there is some way to set the application
Load Order through wlconfig or wldeploy.
I have looked at the WebLogic command reference here: http://e-
docs.bea.com/wls/docs81/admin_ref/ant_tasks.html but don't see anything
promising.
Am I missing something? I'm not too familiar with all the scripting
possibilities so any help would be much appreciated. We are using
WebLogic Server 8.1 SP3 on Windows and UNIX.
ThanksThanks Anilkumar, but I was trying to figure out how to create these using
Ant scripts and NOT the admin console which is what your link seems to be
pointing to.
<akari> wrote in message news:[email protected]..
Hi,
go through the following link,In that you will find information about how
to create execute Queues.
http://e-docs.bea.com/wls/docs81/perform/AppTuning.html#1105201
-----Anilkumar kari -
Hi Folks,
as we got problems on thread deadlocks using JMS, we decided to define our own thread
queue to be used by our application. In the documentation it is described, that you
need to generate all your EJBs with java weblogic.ejbc -dispatchPolicy xyz. By this
you connect your EJB to the named thread queue, which you will need to add to your
config.xml.
What I could find out was, that this works fine for stateful or stateless session
beans, but it looks like message driven beans do ignore this option.
So here are my questions:
Which thread queue is used by message driven beans?
How can I change the thread queue?
Many thanks,
NicoleUnfortunately, MDBs do not currently support running on a separate queue. They run on
the default queue.
-- Rob
Nicole wrote:
Hi Folks,
as we got problems on thread deadlocks using JMS, we decided to define our own thread
queue to be used by our application. In the documentation it is described, that you
need to generate all your EJBs with java weblogic.ejbc -dispatchPolicy xyz. By this
you connect your EJB to the named thread queue, which you will need to add to your
config.xml.
What I could find out was, that this works fine for stateful or stateless session
beans, but it looks like message driven beans do ignore this option.
So here are my questions:
Which thread queue is used by message driven beans?
How can I change the thread queue?
Many thanks,
Nicole
Maybe you are looking for
-
How to order an entire Report by a column with in a group...
Okay, I have a Group Above report, and I want to order the groupings by a column within the group. I am using a lexical parameter in place of the entire ORDER BY statement in the query, giving the user the capability to sort the report by any specifi
-
Physical Model - "New" is grayed out on index creation
Hello, Down in the physical model: I have a table defined with 3 columns. It shows the existing index for the Primary Key for the table. I'm trying to add an index to another column. When I right-click on indexes, the word "New" is grayed out. Can so
-
Configure # of JMS consumers for WebServices over JMS Transport
I'm using WebLogic Server 9.2. My Web Services are using JMS Transport. 10,000 Web Services requests are submitted to the Web Services JMS queue at the same time. Processing time seems a little slow. In the Admin console, monitoring info of the JMS q
-
Adobe Interactive Form Tables with Subtotals
It is possible in WAS 640 to do a table with subtotals using Adobe Designer 6.0? Something as: Key1 Key2 Key3 Value Key1 Key2 Key3 Value Total Key3 Value Key1 Key2 Key4 Value Total Key4 Value Tota
-
IAC HRWPC_FC_EXEC
Hi, I am using the report selection iView (com.sap.pct.erp.mss.reporting) in MSS 600 SP10 (On integrated ITS 640). When I run IAC HRWPC_FC_EXEC the option Menu->List->Export->Local file ... opens the pop-up Save list in file .... There I select Sprea