[JBoss-dev] bug in jbossmq.xml
Hello, some day ago I added a bug report at sourceforge. The first line in file jBoss/conf/default and jboss/conf/tomcat reads ?xml version=1.0 encoding=UTF8? instead of ?xml version=1.0 encoding=UTF-8? because of this jBossmq doesn't start with xerces. The binary distribution version 2.4.0.23 still contains this bug! It's a small typo but newbies may get a bad first impression of jBoss, if it does not run out of the box. Is there a mechanism that someone desides if a update of the binary distibution is nessesary after code changes? Andreas Scott, somebody must have fixed it bevore You checked that the source is ok. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] [ jboss-Feature Requests-453650 ] validation improvements
Feature Requests item #453650, was opened at 2001-08-21 00:58 You can respond by visiting: http://sourceforge.net/tracker/?func=detailatid=376688aid=453650group_id=22866 Category: None Group: None Status: Open Priority: 5 Submitted By: Nobody/Anonymous (nobody) Assigned to: Nobody/Anonymous (nobody) Summary: validation improvements Initial Comment: If a method is declared direct in a home, remote or local interface, but not implemented in the bean class the validator reports a message while deploying the app. But if the method is declared in an super interface the validator check fails. In this case the method setupHomeMappingImpl in class EntityContainer (or a aquivalent) throws a deploeyment exception. But this exception only report the method not the bean or interface name. so one has to check all beans which interfaces implements this super interface. Wy doesen't the validator check also methods inherit form super interfaces? -- You can respond by visiting: http://sourceforge.net/tracker/?func=detailatid=376688aid=453650group_id=22866 ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] SQL function mapping
On Monday, 20. August 2001 19:20, Dain Sundstrom wrote: Thanks, checked it in. -dain snip It is simply two functions position(?2 in substring(?1 from ?3))+?3 Sorry for complaining so late, but it is not as simple as this. position( 'findme' , 'wheretolookfor', 3) should be 0 and not 3 or am I wrong? Therefore use this for PostgreSQL: coalesce( nullif( position( ?1 in substring( ?1 from ?3 ) ), ?3 ) + ?3, 0 ) Marco ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] SQL function mapping
Nope your right. When I first wrote the function I did not have the + ?3 at the end so it worked fine ... Marco Ladermann wrote: On Monday, 20. August 2001 19:20, Dain Sundstrom wrote: Thanks, checked it in. -dain snip It is simply two functions position(?2 in substring(?1 from ?3))+?3 Sorry for complaining so late, but it is not as simple as this. position( 'findme' , 'wheretolookfor', 3) should be 0 and not 3 or am I wrong? Therefore use this for PostgreSQL: coalesce( nullif( position( ?1 in substring( ?1 from ?3 ) ), ?3 ) + ?3, 0 ) Marco ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] Extra ;
Dain in server/src/main/org/jboss/ejb/plugins/cmp/ejbql/Parser.java line 66 you have an extra ';' terminating the return statement. It is causing jikes to complain. Found 1 semantic error compiling /home/dave/jboss/server/src/main/org/jboss/ejb/plugins/cmp/ejbql/Parser.java: [javac] [javac] 66. return (Assembly)completeMatches.iterator().next();; [javac] ^ [javac] *** Error: This statement is unreachable. Could you take it out? Thanks .. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] 2.4.0 has been released
The bundle that I pulled down from SF : http://prdownloads.sourceforge.net/jboss/JBoss-2.4.0.zip contains 2 entire JBoss distribs and hence is twice the size of a normal download... I diffed them. lib/ext/jboss.jar differed between them. the one with the BETA suffix had a load of extra dirs Is this intentional - I was confisued as to which one I should use to build JBoss/Jetty. Jules --- Scott M Stark [EMAIL PROTECTED] wrote: The initial release of the 2.4.0 binary has been released and is available at sourceforge. Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] bug in jbossmq.xml
This does not exist in the 2.4.0 final release made yesterday. - Original Message - From: Schouten, Andreas [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, August 21, 2001 12:06 AM Subject: [JBoss-dev] bug in jbossmq.xml Hello, some day ago I added a bug report at sourceforge. The first line in file jBoss/conf/default and jboss/conf/tomcat reads ?xml version=1.0 encoding=UTF8? instead of ?xml version=1.0 encoding=UTF-8? because of this jBossmq doesn't start with xerces. The binary distribution version 2.4.0.23 still contains this bug! It's a small typo but newbies may get a bad first impression of jBoss, if it does not run out of the box. Is there a mechanism that someone desides if a update of the binary distibution is nessesary after code changes? Andreas Scott, somebody must have fixed it bevore You checked that the source is ok. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] 2.4.0 has been released
Not intended. I gues the dist target doesn't clean out any existing dirs before creating the archive. I'll repackage it. - Original Message - From: Julian Gosnell [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, August 21, 2001 4:56 AM Subject: Re: [JBoss-dev] 2.4.0 has been released The bundle that I pulled down from SF : http://prdownloads.sourceforge.net/jboss/JBoss-2.4.0.zip contains 2 entire JBoss distribs and hence is twice the size of a normal download... I diffed them. lib/ext/jboss.jar differed between them. the one with the BETA suffix had a load of extra dirs Is this intentional - I was confisued as to which one I should use to build JBoss/Jetty. Jules --- Scott M Stark [EMAIL PROTECTED] wrote: The initial release of the 2.4.0 binary has been released and is available at sourceforge. Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jboss/src/etc/conf/default mail.properties
User: starksm Date: 01/08/21 08:37:52 Modified:src/etc/conf/default Tag: Branch_2_4 mail.properties Log: Don't use Simone's mail info as the default Revision ChangesPath No revision No revision 1.2.4.1 +12 -6 jboss/src/etc/conf/default/mail.properties Index: mail.properties === RCS file: /cvsroot/jboss/jboss/src/etc/conf/default/mail.properties,v retrieving revision 1.2 retrieving revision 1.2.4.1 diff -u -r1.2 -r1.2.4.1 --- mail.properties 2001/04/20 22:55:30 1.2 +++ mail.properties 2001/08/21 15:37:51 1.2.4.1 @@ -1,7 +1,13 @@ -mail.store.protocol = pop3 +# Change to your mail server prototocol +mail.store.protocol = pop3 mail.transport.protocol = smtp -mail.user= biorn_steedom -mail.pop3.host = pop.mail.yahoo.com -mail.smtp.host = smtp.mail.yahoo.com -mail.from= [EMAIL PROTECTED] -mail.debug = false +# Change to the user who will receive mail +mail.user = to_user +# Change to the mail server +mail.pop3.host = pop.dot.com +# Change to the SMTP gateway server +mail.smtp.host = smtp.dot.com +# Change to the address mail will be from +mail.from = [EMAIL PROTECTED] +# Enable debugging output from the javamail classes +mail.debug = false ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jboss/src/etc/conf/default mail.properties
User: starksm Date: 01/08/21 08:57:19 Modified:src/etc/conf/default mail.properties Log: Change the default properties from Simone's to values that must be customized Revision ChangesPath 1.4 +13 -7 jboss/src/etc/conf/default/mail.properties Index: mail.properties === RCS file: /cvsroot/jboss/jboss/src/etc/conf/default/mail.properties,v retrieving revision 1.3 retrieving revision 1.4 diff -u -r1.3 -r1.4 --- mail.properties 2001/08/11 21:32:45 1.3 +++ mail.properties 2001/08/21 15:57:18 1.4 @@ -4,12 +4,18 @@ ## ## ### == ### -### $Id: mail.properties,v 1.3 2001/08/11 21:32:45 user57 Exp $ ### +### $Id: mail.properties,v 1.4 2001/08/21 15:57:18 starksm Exp $ ### -mail.store.protocol = pop3 +# Change to your mail server prototocol +mail.store.protocol = pop3 mail.transport.protocol = smtp -mail.user = biorn_steedom -mail.pop3.host = pop.mail.yahoo.com -mail.smtp.host = smtp.mail.yahoo.com -mail.from = [EMAIL PROTECTED] -mail.debug = false +# Change to the user who will receive mail +mail.user = to_user +# Change to the mail server +mail.pop3.host = pop.dot.com +# Change to the SMTP gateway server +mail.smtp.host = smtp.dot.com +# Change to the address mail will be from +mail.from = [EMAIL PROTECTED] +# Enable debugging output from the javamail classes +mail.debug = false ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: newsite binary.jsp
User: starksm Date: 01/08/21 09:13:33 Modified:.binary.jsp Log: Update the size and dates of the 2.4.0 release Revision ChangesPath 1.9 +4 -4 newsite/binary.jsp Index: binary.jsp === RCS file: /cvsroot/jboss/newsite/binary.jsp,v retrieving revision 1.8 retrieving revision 1.9 diff -u -r1.8 -r1.9 --- binary.jsp2001/08/21 01:06:41 1.8 +++ binary.jsp2001/08/21 16:13:33 1.9 @@ -36,14 +36,14 @@ tr tda class=link href=http://prdownloads.sourceforge.net/jboss/JBoss-2.4.0.zip;JBoss-2.4.0.zip/a/td -td align=centerp class=text16233873/td -td align=centerp class=textAugust 20, 2001/td +td align=centerp class=text8109497/td +td align=centerp class=textAugust 21, 2001/td /tr tr tda class=link href=http://prdownloads.sourceforge.net/jboss/JBoss-2.4.0_Tomcat-3.2.3.zip;JBoss-2.4.0_Tomcat-3.2.3.zip/a/td -td align=centerp class=text11414416/td -td align=centerp class=textAugust 20, 2001/td +td align=centerp class=text11415568/td +td align=centerp class=textAugust 21, 2001/td /tr tr ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] 2.4.0 has been released
The new package has been placed on sourceforge. This should appear soon under the Files section, but is available now via http://prdownloads.sourceforge.net/jboss/JBoss-2.4.0.zip - Original Message - From: Scott M Stark [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, August 21, 2001 8:22 AM Subject: Re: [JBoss-dev] 2.4.0 has been released Not intended. I gues the dist target doesn't clean out any existing dirs before creating the archive. I'll repackage it. - Original Message - From: Julian Gosnell [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, August 21, 2001 4:56 AM Subject: Re: [JBoss-dev] 2.4.0 has been released The bundle that I pulled down from SF : http://prdownloads.sourceforge.net/jboss/JBoss-2.4.0.zip contains 2 entire JBoss distribs and hence is twice the size of a normal download... I diffed them. lib/ext/jboss.jar differed between them. the one with the BETA suffix had a load of extra dirs Is this intentional - I was confisued as to which one I should use to build JBoss/Jetty. Jules --- Scott M Stark [EMAIL PROTECTED] wrote: The initial release of the 2.4.0 binary has been released and is available at sourceforge. Do You Yahoo!? Get your free @yahoo.co.uk address at http://mail.yahoo.co.uk or your free @yahoo.ie address at http://mail.yahoo.ie ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore()
Jason, I'm going to combine multiple email replies into one. Please switch back to QueuedPessimisticEJBLock. We're trying to make this the default for JBoss since it scales better. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Jason Dillon Sent: Monday, August 20, 2001 7:06 PM To: [EMAIL PROTECTED] Subject: Re: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore() I am using commit option B, since it might be possible that another process is using the database... though I am 100% sure that I have sole access to the database I am using. You probably already know this but ejbLoad will get called whenever the entity becomes part of the transaction If I run the first producer with the first MDB offline, then connect the MDB everything seems to be fine. I have run a few tests through in this manner and I have not had problem yet, where as if I try to have the both up together it will work only once after all nodes have been started (usually) and will fail most times after that. I should be able to have more than one thread update a single entity at the same time (or rather ordered as the container dictates) right? By more than one thread you mean more than one transaction? Yes, you can have more than one transaction, but remember, this is pessimistic locking. The entity will be locked until the transaction commits. All other accesses to the entity will be blocked until the entity is committed by the first transaction. so you see: 1. setModified() 2. ejbLoad So there is no ejbStore() in between setModified() and ejbLoad()? Correct. I also very rarly a javax.ejb.DuplicateKeyException, which happend to just pop up as I was running these tests. I assume you're using commit option 'A'? Nope, option 'B'. isModified only get's called before a storeEntity(ejbStore()) It does not get called before an ejbLoad(). storeEntity will get called on a transaction commit. Also when a finder or remove method is called, but only beans within the transaction will get stored in this case. -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Jason Dillon Sent: Monday, August 20, 2001 7:34 PM To: [EMAIL PROTECTED] Subject: RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore() On Mon, 20 Aug 2001, Bill Burke wrote: A thing to note. Whenever a finder or remove is called within a transaction, every entity of every type that is part of the transaction is synchronized with the database. Could this be your problem? The (5) MDB uses findByPrimaryKey to locate all of the beans (a, b, c d). So, if there is an entity and modified by one thread under one tx, and another thread/tx does a findby, the first thread will loose its data due to a ejbLoad from the second? Not at all. Remember. Pessimistic locking! Once an entity becomes part of a tx, all other threads/transactions are blocked until the tx commits. This is essentially what I am seeing, though I can not really tell that this I just found a bug in QueuedPessimisticEJBLock that may be causing this problem. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] Clustering and SOAP
Is there an archive where the results of the jboss discussion on how clustering will be implemented (and when) in JBoss? I assume that the clustering is for horizontal load balancing, fail over, etc. and that vertical mechanism are already done. Is there a summary of the discussion as to why a separate SOAP implementation was done (zoap) rather than use the apache SOAP? Thanks. RME ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
AW: [JBoss-dev] Clustering and SOAP
-Ursprüngliche Nachricht- Von: Richard Emberson [mailto:[EMAIL PROTECTED]] Gesendet: Dienstag, 21. August 2001 18:35 An: [EMAIL PROTECTED] Betreff: [JBoss-dev] Clustering and SOAP Is there a summary of the discussion as to why a separate SOAP implementation was done (zoap) rather than use the apache SOAP? Cos apache-SOAP (still) sucks, IMHO. Wait. Polemics off. At the time we started with ZOAP (must have been late 2000, or so, cant remember the exact month), ApacheSOAP was hardly able to serialize serious data structures besides the strings and numerics from the StockTicker demo (not to mention the performance issues due to using DOM). Due to a lacking acceptance of ZOAP, however, we have decided to abandon the ZOAP effort in favor to Apache/Axis/whatever. Unfortunately, there is not yet enough critical mass to form a new subproject, the exact target (Invocation layer, JMX adaptor, JMS integration) is also not clear. But at least discussion is underway ... CGJ ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] Clustering and SOAP
I'm not sure about the clustering stuff, but ZOAP was developed before Apache SOAP was very usable. Since apache SOAP has become more functional ZOAP is no longer under development. Search for posts by Dr. Jung. On 2001.08.21 12:34:58 -0400 Richard Emberson wrote: Is there an archive where the results of the jboss discussion on how clustering will be implemented (and when) in JBoss? I assume that the clustering is for horizontal load balancing, fail over, etc. and that vertical mechanism are already done. Is there a summary of the discussion as to why a separate SOAP implementation was done (zoap) rather than use the apache SOAP? Thanks. RME ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] bug in jbossmq.xml
To fully fix that bug, I recently changed the: jbossmq\src\main\org\jbossmq\xml\XElement.java file. JBossMQ was incorrectly rewriting the jbossmq.xml with the UTF8 string. I commited the change to the 2.4 branch.. But I don't know if it was rebuilt and included in the jboss libs for the final release. Regards, Hiram From: Scott M Stark [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: [JBoss-dev] bug in jbossmq.xml Date: Tue, 21 Aug 2001 07:32:53 -0700 This does not exist in the 2.4.0 final release made yesterday. - Original Message - From: Schouten, Andreas [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Tuesday, August 21, 2001 12:06 AM Subject: [JBoss-dev] bug in jbossmq.xml Hello, some day ago I added a bug report at sourceforge. The first line in file jBoss/conf/default and jboss/conf/tomcat reads ?xml version=1.0 encoding=UTF8? instead of ?xml version=1.0 encoding=UTF-8? because of this jBossmq doesn't start with xerces. The binary distribution version 2.4.0.23 still contains this bug! It's a small typo but newbies may get a bad first impression of jBoss, if it does not run out of the box. Is there a mechanism that someone desides if a update of the binary distibution is nessesary after code changes? Andreas Scott, somebody must have fixed it bevore You checked that the source is ok. ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development _ Get your FREE download of MSN Explorer at http://explorer.msn.com/intl.asp ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jboss/src/main/org/jboss/ejb/plugins/lock QueuedPessimisticEJBLock.java
User: patriot1burke
Date: 01/08/21 11:35:51
Modified:src/main/org/jboss/ejb/plugins/lock
QueuedPessimisticEJBLock.java
Log:
Needed to set tx after obtaining lock
Revision ChangesPath
1.3 +12 -9
jboss/src/main/org/jboss/ejb/plugins/lock/QueuedPessimisticEJBLock.java
Index: QueuedPessimisticEJBLock.java
===
RCS file:
/cvsroot/jboss/jboss/src/main/org/jboss/ejb/plugins/lock/QueuedPessimisticEJBLock.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- QueuedPessimisticEJBLock.java 2001/08/08 22:13:00 1.2
+++ QueuedPessimisticEJBLock.java 2001/08/21 18:35:51 1.3
@@ -42,7 +42,7 @@
* @author a href=[EMAIL PROTECTED]Marc Fleury/a
* @author a href=[EMAIL PROTECTED]Bill Burke/a
*
- * @version $Revision: 1.2 $
+ * @version $Revision: 1.3 $
*
* pbRevisions:/bbr
* pb2001/08/03: billb/b
@@ -156,10 +156,10 @@
protected boolean doSchedule(MethodInvocation mi)
throws Exception
{
- this.sync();
boolean wasThreadScheduled = false;
Transaction miTx = mi.getTransaction();
boolean trace = log.isTraceEnabled();
+ this.sync();
try
{
if( trace ) log.trace(Begin schedule, key=+mi.getId());
@@ -249,9 +249,9 @@
// We loop here until either until success or until transaction timeout
// If we get out of the loop successfully, we can successfully
// set the transaction on this puppy.
- while (tx != null
+ while (this.tx != null
// And are we trying to enter with another transaction?
- !tx.equals(miTx))
+ !this.tx.equals(miTx))
{
wasScheduled = true;
// That's no good, only one transaction per context
@@ -260,7 +260,7 @@
TxLock txLock = getTxLock(miTx);
- if( trace ) log.trace(Begin wait on Tx=+tx);
+ if( trace ) log.trace(Begin wait on Tx=+this.tx);
// And lock the threads on the lock corresponding to the Tx in MI
synchronized(txLock)
@@ -274,7 +274,7 @@
this.sync();
- if( trace ) log.trace(End wait on TxLock=+tx);
+ if( trace ) log.trace(End wait on TxLock=+this.tx);
if (isTxExpired(miTx))
{
log.error(Thread.currentThread() + Saw rolled back tx=+miTx+ waiting
for txLock
@@ -289,7 +289,7 @@
txLocks.remove(txLock);
txWaitQueue.remove(txLock);
}
-else if (tx != null tx.equals(miTx))
+else if (this.tx != null tx.equals(miTx))
{
// We're not qu
nextTransaction();
@@ -297,6 +297,9 @@
throw new RuntimeException(Transaction marked for rollback, possibly a
timeout);
}
} // end while(tx!=miTx)
+
+ // If we get here, this means that we have the txlock
+ this.tx = miTx;
return wasScheduled;
}
@@ -405,7 +408,7 @@
if (numMethodLocks == 0)
{
synchronized(methodLock) {methodLock.notify();}
- }
+ }
}
/*
@@ -438,7 +441,7 @@
// saveStackTrace(removing bean lock and it has tx's in
QUEUE!***);
throw new IllegalStateException(removing bean lock and it has tx's in
QUEUE!);
}
- else if (refs == 0 tx != null)
+ else if (refs == 0 this.tx != null)
{
// saveStackTrace(removing bean lock and it has tx set!***);
throw new IllegalStateException(removing bean lock and it has tx set!);
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore()
I'm going to combine multiple email replies into one. Please switch back to QueuedPessimisticEJBLock. We're trying to make this the default for JBoss since it scales better. Ok. I was not sure if one was more stable than the other. I am using commit option B, since it might be possible that another process is using the database... though I am 100% sure that I have sole access to the database I am using. You probably already know this but ejbLoad will get called whenever the entity becomes part of the transaction I figured that much, but I was not really sure. If I run the first producer with the first MDB offline, then connect the MDB everything seems to be fine. I have run a few tests through in this manner and I have not had problem yet, where as if I try to have the both up together it will work only once after all nodes have been started (usually) and will fail most times after that. I should be able to have more than one thread update a single entity at the same time (or rather ordered as the container dictates) right? By more than one thread you mean more than one transaction? Yes, you can have more than one transaction, but remember, this is pessimistic locking. The entity will be locked until the transaction commits. All other accesses to the entity will be blocked until the entity is committed by the first transaction. That is not the behavior that I was seeing, perhaps your change to Queued* fixed that. Not sure at the moment. I think I did find one rather large application error, where the last MDB updated an entity, and then tried to remove it (and all other related entities) at the end, all inside of one tx. I made a dumb ass mistake and changed the method which removed these to be NotSupported, which was causing a TX timeout waiting for the calling TX to complete before anything could be removed... Since the last update never commits, the database always looks inconsistent. Any ways, I think that this was the problem that I was seeing with the Simple* lock. Not at all. Remember. Pessimistic locking! Once an entity becomes part of a tx, all other threads/transactions are blocked until the tx commits. This is essentially what I am seeing, though I can not really tell that this I just found a bug in QueuedPessimisticEJBLock that may be causing this problem. Cool. I will let you know if this fixes the problem I am seeing. --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore()
-Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Jason Dillon Sent: Tuesday, August 21, 2001 4:36 PM To: [EMAIL PROTECTED] Subject: RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore() I'm going to combine multiple email replies into one. Please switch back to QueuedPessimisticEJBLock. We're trying to make this the default for JBoss since it scales better. Ok. I was not sure if one was more stable than the other. I am using commit option B, since it might be possible that another process is using the database... though I am 100% sure that I have sole access to the database I am using. You probably already know this but ejbLoad will get called whenever the entity becomes part of the transaction I figured that much, but I was not really sure. If I run the first producer with the first MDB offline, then connect the MDB everything seems to be fine. I have run a few tests through in this manner and I have not had problem yet, where as if I try to have the both up together it will work only once after all nodes have been started (usually) and will fail most times after that. I should be able to have more than one thread update a single entity at the same time (or rather ordered as the container dictates) right? By more than one thread you mean more than one transaction? Yes, you can have more than one transaction, but remember, this is pessimistic locking. The entity will be locked until the transaction commits. All other accesses to the entity will be blocked until the entity is committed by the first transaction. That is not the behavior that I was seeing, perhaps your change to Queued* fixed that. Not sure at the moment. I think I did find one rather large The Queued* bug fix should show the behaviour I talked about above. Let me know. Thanks, Bill ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore()
I have not seen this problem again... so far. I am running larger tests to confirm. --jason On Tue, 21 Aug 2001, Bill Burke wrote: -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Jason Dillon Sent: Tuesday, August 21, 2001 4:36 PM To: [EMAIL PROTECTED] Subject: RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore() I'm going to combine multiple email replies into one. Please switch back to QueuedPessimisticEJBLock. We're trying to make this the default for JBoss since it scales better. Ok. I was not sure if one was more stable than the other. I am using commit option B, since it might be possible that another process is using the database... though I am 100% sure that I have sole access to the database I am using. You probably already know this but ejbLoad will get called whenever the entity becomes part of the transaction I figured that much, but I was not really sure. If I run the first producer with the first MDB offline, then connect the MDB everything seems to be fine. I have run a few tests through in this manner and I have not had problem yet, where as if I try to have the both up together it will work only once after all nodes have been started (usually) and will fail most times after that. I should be able to have more than one thread update a single entity at the same time (or rather ordered as the container dictates) right? By more than one thread you mean more than one transaction? Yes, you can have more than one transaction, but remember, this is pessimistic locking. The entity will be locked until the transaction commits. All other accesses to the entity will be blocked until the entity is committed by the first transaction. That is not the behavior that I was seeing, perhaps your change to Queued* fixed that. Not sure at the moment. I think I did find one rather large The Queued* bug fix should show the behaviour I talked about above. Let me know. Thanks, Bill ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore()
I have pushed close to 600x3 transactions through our system and have not run into any problems yet. Once I get to 1000x3 I will know for sure. Looks really good so far! --jason On Tue, 21 Aug 2001, Jason Dillon wrote: I have not seen this problem again... so far. I am running larger tests to confirm. --jason On Tue, 21 Aug 2001, Bill Burke wrote: -Original Message- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]On Behalf Of Jason Dillon Sent: Tuesday, August 21, 2001 4:36 PM To: [EMAIL PROTECTED] Subject: RE: [JBoss-dev] ejbLoad() on a modified bean w/o ejbStore() I'm going to combine multiple email replies into one. Please switch back to QueuedPessimisticEJBLock. We're trying to make this the default for JBoss since it scales better. Ok. I was not sure if one was more stable than the other. I am using commit option B, since it might be possible that another process is using the database... though I am 100% sure that I have sole access to the database I am using. You probably already know this but ejbLoad will get called whenever the entity becomes part of the transaction I figured that much, but I was not really sure. If I run the first producer with the first MDB offline, then connect the MDB everything seems to be fine. I have run a few tests through in this manner and I have not had problem yet, where as if I try to have the both up together it will work only once after all nodes have been started (usually) and will fail most times after that. I should be able to have more than one thread update a single entity at the same time (or rather ordered as the container dictates) right? By more than one thread you mean more than one transaction? Yes, you can have more than one transaction, but remember, this is pessimistic locking. The entity will be locked until the transaction commits. All other accesses to the entity will be blocked until the entity is committed by the first transaction. That is not the behavior that I was seeing, perhaps your change to Queued* fixed that. Not sure at the moment. I think I did find one rather large The Queued* bug fix should show the behaviour I talked about above. Let me know. Thanks, Bill ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] manual build
Hi, Is there anybody trying to make the manual build run correctly? I give it a (painful) try. I succeed to build a part of html and printable-html (hum simply the preface,intro, cmp and jaws for the moment) I choose to move to ant 1.4 that use JAXP 1.1 to avoid problems with style, xsl, ... I needed to change the jboss-all/build/build.bat to include sun/jaxp and ant/lib/ext jars. The manual build.xml has also changed. In jbossdocs.xml I have to reference the _entity_ using full path Because of these (ant 1.4 is beta and I did not check compliance with other modules), I don't know if it is a good idea to commit now. What do you think ? Create a branch until it works ? The problems I have now: printable-html : Some javax.xml.transform.TransformerException when I added more chapter xml files to jbossdocs.xml. printable-html : the CSS is nt included in generated html. printable-html : the file is jbossdocshtml (the . is missing) html: it fails everytime with a javax.xml.transform.TransformerException: java.lang.ArrayIndexOutOfBoundsException at the end of the build. Vincent. _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] jbossmq load balancing across vms
Hello again, I am running into a problem with the latest JBossMQ wrt using it to load balance across machines. I have one machine which acts as a message router, running JBossMQ. All other machines use it for their JMSProvider. When I have more than one MDB attached to a single queue, and a user-thread adding messages over and over, it looks like only one of the MDB's will ever read any messages. I do not remember seeing this before. It looks like when I startup a new MDB that 50% of the time it will take over and the others will remain idle, or it will simply never read any new messages. I have each of these MDB setup with: MaximumSize1/MaximumSize MaxMessages1/MaxMessages Which could be causing problem... but honestly I don't really know what these two setting actually do. I wanted to have these deployed MDB's to only process one message at a time, which this seems to accomplish most of the time. Any ways, I am concerended about how JBossMQ is pushing messages over to consumers. I mentioned before that if one producer added 1k messages to a queue, then one MDB was attached which would suck up all of the messages. When a second MDB was attached it would not have anything todo. Hiram mentioned something about this, but I am not really sure what the plans were to fix this. I would like to get this fixed soon, as I am fairly dependent on this behavior. From my perspective, there are two major issues: 1) JBossMQ does not seem to trade off between consumers (ie. round-robin or whatever). 2) JBossMQ does not have any configurable limits on how messages are pushed from the server to clients/consumers. I understand why JBossMQ wants to push over bulk messages to consumers, but that is only really useful in some applications. In my case, I would rather that it pushed over smallers chunks (or even 1 msg at a time) to clients so that I could utilize the cluster of machines that I have to distribute the processing load. So, to the JBossMQ folks, does this sound like it would be easy to implement? Do you have any suggestions on how it might be done (quick fix and full blown god JMS provider version)? I am willing to look into this more, but I could use some pointers on which bits need to be tweaked. Any ideas? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmq/src/main/org/jboss/mq/cluster/jms ClusterTopicSession.java ClusterTopicSubscriber.java
User: user57
Date: 01/08/21 16:16:18
Modified:src/main/org/jboss/mq/cluster/jms ClusterTopicSession.java
ClusterTopicSubscriber.java
Log:
o removed org.jboss.mq.selectors.Selector from import, was not being used.
Revision ChangesPath
1.3 +4 -3
jbossmq/src/main/org/jboss/mq/cluster/jms/ClusterTopicSession.java
Index: ClusterTopicSession.java
===
RCS file:
/cvsroot/jboss/jbossmq/src/main/org/jboss/mq/cluster/jms/ClusterTopicSession.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- ClusterTopicSession.java 2001/08/17 03:04:03 1.2
+++ ClusterTopicSession.java 2001/08/21 23:16:18 1.3
@@ -36,16 +36,17 @@
import org.jboss.mq.SpyObjectMessage;
import org.jboss.mq.SpyStreamMessage;
import org.jboss.mq.SpyTextMessage;
-import org.jboss.mq.selectors.Selector;
/**
* This class implements javax.jms.TopicSession
*
* @author Hiram Chirino ([EMAIL PROTECTED])
* @createdAugust 16, 2001
- * @version$Revision: 1.2 $
+ * @version$Revision: 1.3 $
*/
-public class ClusterTopicSession implements TopicSession {
+public class ClusterTopicSession
+ implements TopicSession
+{
//MessageConsumers created by this session
protected HashSet consumers = new HashSet();
1.3 +5 -6
jbossmq/src/main/org/jboss/mq/cluster/jms/ClusterTopicSubscriber.java
Index: ClusterTopicSubscriber.java
===
RCS file:
/cvsroot/jboss/jbossmq/src/main/org/jboss/mq/cluster/jms/ClusterTopicSubscriber.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- ClusterTopicSubscriber.java 2001/08/17 03:04:03 1.2
+++ ClusterTopicSubscriber.java 2001/08/21 23:16:18 1.3
@@ -16,17 +16,16 @@
import org.jboss.mq.SpyMessage;
import org.jboss.mq.cluster.transport.NodeId;
-import org.jboss.mq.selectors.Selector;
-
/**
- * This class implements javax.jms.TopicSubscriber
+ * This class implements ttjavax.jms.TopicSubscriber/tt.
*
* @author Hiram Chirino ([EMAIL PROTECTED])
* @createdAugust 16, 2001
- * @version$Revision: 1.2 $
+ * @version$Revision: 1.3 $
*/
-public class ClusterTopicSubscriber implements TopicSubscriber {
-
+public class ClusterTopicSubscriber
+ implements TopicSubscriber
+{
// The topic I registered
Topictopic;
//Am I closed ?
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmq/src/main/org/jboss/mq SpyMessageConsumer.java SpyTopicSession.java SpyTopicSubscriber.java
User: user57
Date: 01/08/21 16:16:17
Modified:src/main/org/jboss/mq SpyMessageConsumer.java
SpyTopicSession.java SpyTopicSubscriber.java
Log:
o removed org.jboss.mq.selectors.Selector from import, was not being used.
Revision ChangesPath
1.7 +5 -6 jbossmq/src/main/org/jboss/mq/SpyMessageConsumer.java
Index: SpyMessageConsumer.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/SpyMessageConsumer.java,v
retrieving revision 1.6
retrieving revision 1.7
diff -u -r1.6 -r1.7
--- SpyMessageConsumer.java 2001/08/17 03:04:01 1.6
+++ SpyMessageConsumer.java 2001/08/21 23:16:17 1.7
@@ -16,19 +16,18 @@
import javax.jms.MessageListener;
import javax.jms.Session;
-import org.jboss.mq.selectors.Selector;
-
/**
- * This class implements javax.jms.MessageConsumer
+ * This class implements ttjavax.jms.MessageConsumer/tt.
*
* @author Norbert Lataille ([EMAIL PROTECTED])
* @author Hiram Chirino ([EMAIL PROTECTED])
* @author David Maplesden ([EMAIL PROTECTED])
* @createdAugust 16, 2001
- * @version$Revision: 1.6 $
+ * @version$Revision: 1.7 $
*/
-public class SpyMessageConsumer implements MessageConsumer, SpyConsumer, Runnable {
-
+public class SpyMessageConsumer
+ implements MessageConsumer, SpyConsumer, Runnable
+{
//Link to my session
public SpySession session;
// The subscription structure should be fill out by the decendent
1.3 +5 -6 jbossmq/src/main/org/jboss/mq/SpyTopicSession.java
Index: SpyTopicSession.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/SpyTopicSession.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- SpyTopicSession.java 2001/08/17 03:04:01 1.2
+++ SpyTopicSession.java 2001/08/21 23:16:17 1.3
@@ -24,20 +24,19 @@
import javax.jms.TopicSubscriber;
import javax.jms.XATopicSession;
-import org.jboss.mq.selectors.Selector;
-
/**
- * This class implements javax.jms.TopicSession and javax.jms.XATopicSession
+ * This class implements ttjavax.jms.TopicSession/tt and
+ * ttjavax.jms.XATopicSession/tt.
*
* @author Norbert Lataille ([EMAIL PROTECTED])
* @author Hiram Chirino ([EMAIL PROTECTED])
* @createdAugust 16, 2001
- * @version$Revision: 1.2 $
+ * @version$Revision: 1.3 $
*/
public class SpyTopicSession
extends SpySession
- implements TopicSession, javax.jms.XATopicSession {
-
+ implements TopicSession, XATopicSession
+{
SpyTopicSession( Connection myConnection, boolean transacted, int
acknowledgeMode ) {
this( myConnection, transacted, acknowledgeMode, false );
}
1.3 +6 -6 jbossmq/src/main/org/jboss/mq/SpyTopicSubscriber.java
Index: SpyTopicSubscriber.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/SpyTopicSubscriber.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- SpyTopicSubscriber.java 2001/08/17 03:04:01 1.2
+++ SpyTopicSubscriber.java 2001/08/21 23:16:17 1.3
@@ -5,6 +5,7 @@
* See terms of license at gnu.org.
*/
package org.jboss.mq;
+
import javax.jms.JMSException;
import javax.jms.Message;
import javax.jms.MessageListener;
@@ -12,22 +13,21 @@
import javax.jms.TopicSubscriber;
-import org.jboss.mq.selectors.Selector;
-
/**
- * This class implements javax.jms.TopicSubscriber
+ * This class implements ttjavax.jms.TopicSubscriber/tt.
*
* @author Norbert Lataille ([EMAIL PROTECTED])
* @author Hiram Chirino ([EMAIL PROTECTED])
* @createdAugust 16, 2001
- * @version$Revision: 1.2 $
+ * @version$Revision: 1.3 $
*/
public class SpyTopicSubscriber
extends SpyMessageConsumer
- implements TopicSubscriber {
+ implements TopicSubscriber
+{
// Attributes
- //The topic I registered
+ /** The registered topic. */
private SpyTopic topic;
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmq/src/main/org/jboss/mq Subscription.java
User: user57
Date: 01/08/21 16:21:51
Modified:src/main/org/jboss/mq Subscription.java
Log:
o changed comments to javadocs
Revision ChangesPath
1.3 +32 -24jbossmq/src/main/org/jboss/mq/Subscription.java
Index: Subscription.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/Subscription.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- Subscription.java 2001/08/17 03:04:01 1.2
+++ Subscription.java 2001/08/21 23:21:51 1.3
@@ -8,63 +8,71 @@
import java.io.Serializable;
+import javax.jms.JMSException;
+
import org.jboss.mq.selectors.Selector;
/**
- * This class contians all the data needed to for a the provider to to
- * determine if a message can be routed to a consumer.
+ * This class contains all the data needed to for a the provider to to
+ * determine if a message can be routed to a consumer.
*
* @author Hiram Chirino ([EMAIL PROTECTED])
* @author David Maplesden ([EMAIL PROTECTED])
* @createdAugust 16, 2001
- * @version$Revision: 1.2 $
+ * @version$Revision: 1.3 $
*/
public class Subscription
- implements Serializable {
- // This gets set to a unique value at the SpyConnection
+ implements Serializable
+{
+ /** This gets set to a unique value at the SpyConnection. */
public int subscriptionId;
- // the queue we want to subscribe to
+
+ /** The queue we want to subscribe to. */
public SpyDestination destination;
- // the selector which will filter out messages
+
+ /** The selector which will filter out messages. */
public StringmessageSelector;
- // Should this message destroy the subscription?
+
+ /** Should this message destroy the subscription? */
public boolean destroyDurableSubscription;
- // Topics might not want locally produced messages
+ /** Topics might not want locally produced messages. */
public boolean noLocal;
// Transient Values
public transient Selector selector;
public transient ConnectionToken dc;
- public transient Object clientConsumer = null;
-
- // Determines the consumer would accept the message.
- public Selector getSelector()
- throws javax.jms.JMSException {
+ public transient Object clientConsumer; // = null;
- if ( messageSelector == null ) {
+ /**
+* Determines the consumer would accept the message.
+*/
+ public Selector getSelector() throws JMSException {
+ if (messageSelector == null) {
return null;
}
- if ( selector == null ) {
- selector = new Selector( messageSelector );
+ if (selector == null) {
+ selector = new Selector(messageSelector);
}
return selector;
}
- // Determines the consumer would accept the message.
- public boolean accepts( SpyMessage message )
- throws javax.jms.JMSException {
- if ( message.getJMSDestination() instanceof SpyTopic ) {
+ /**
+* Determines the consumer would accept the message.
+*/
+ public boolean accepts(SpyMessage message) throws JMSException {
+ if (message.getJMSDestination() instanceof SpyTopic) {
// In the Topic case we allways deliver unless we have a noLocal
- if ( noLocal message.producerClientId.equals( dc.getClientID() ) ) {
+ if (noLocal message.producerClientId.equals(dc.getClientID())) {
return false;
}
}
+
Selector ms = getSelector();
- if ( ms != null ) {
- if ( !ms.test( message ) ) {
+ if (ms != null) {
+ if (!ms.test(message)) {
return false;
}
}
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] jbossmq load balancing across vms
Hi Jason, The way we implemented the message delivery this behaviour could happen, but only when messages are arriving more slowly than they are being requested. This is because the set of receivers waiting for messages is a HashSet so that if a receiver sends a request for a message more than once (which happens when using a receive with a timeout) it won't get added more than once. What this means is that if you have multiple receivers waiting for a message from a queue and a message arrives then it will be immediately delivered to receiver A. Receiver A is now removed from the HashSet, however if receiver A finishes its processing and requests another message before another message is sent to the queue it will be added back to the HashSet. When the next message arrives set will again be sent to receiver A because in the HashSets ordering receiver A is clearly ahead of receiver B. We didn't see any problem with this at the time, under heavy client processing loads where messages are waiting on the queue as requests are made then the messages will be shared out amongst the clients. When there is light loads then the messages will tend to be delivered to the same client. We could implement an ordering on the server-side but I still don't know if it is required. Remember this only happens when your clients are going fast enough that they actually have to wait for messages to arrive. If messages are being produced faster than the clients can process them then they will be shared out amongst all the clients. Cheers David -Original Message- From: Jason Dillon [mailto:[EMAIL PROTECTED]] Sent: Wednesday, August 22, 2001 11:06 AM To: [EMAIL PROTECTED] Subject: [JBoss-dev] jbossmq load balancing across vms Hello again, I am running into a problem with the latest JBossMQ wrt using it to load balance across machines. I have one machine which acts as a message router, running JBossMQ. All other machines use it for their JMSProvider. When I have more than one MDB attached to a single queue, and a user-thread adding messages over and over, it looks like only one of the MDB's will ever read any messages. I do not remember seeing this before. It looks like when I startup a new MDB that 50% of the time it will take over and the others will remain idle, or it will simply never read any new messages. I have each of these MDB setup with: MaximumSize1/MaximumSize MaxMessages1/MaxMessages Which could be causing problem... but honestly I don't really know what these two setting actually do. I wanted to have these deployed MDB's to only process one message at a time, which this seems to accomplish most of the time. Any ways, I am concerended about how JBossMQ is pushing messages over to consumers. I mentioned before that if one producer added 1k messages to a queue, then one MDB was attached which would suck up all of the messages. When a second MDB was attached it would not have anything todo. Hiram mentioned something about this, but I am not really sure what the plans were to fix this. I would like to get this fixed soon, as I am fairly dependent on this behavior. From my perspective, there are two major issues: 1) JBossMQ does not seem to trade off between consumers (ie. round-robin or whatever). 2) JBossMQ does not have any configurable limits on how messages are pushed from the server to clients/consumers. I understand why JBossMQ wants to push over bulk messages to consumers, but that is only really useful in some applications. In my case, I would rather that it pushed over smallers chunks (or even 1 msg at a time) to clients so that I could utilize the cluster of machines that I have to distribute the processing load. So, to the JBossMQ folks, does this sound like it would be easy to implement? Do you have any suggestions on how it might be done (quick fix and full blown god JMS provider version)? I am willing to look into this more, but I could use some pointers on which bits need to be tweaked. Any ideas? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jboss-j2ee/src/main/javax/jms JMSException.java
User: user57
Date: 01/08/21 16:47:10
Modified:src/main/javax/jms JMSException.java
Log:
o fixed setLinkedException sig to be serial compatible with JMS 1.0.2
(needed to be synchronized and have its param name as 'ex' instead of
'linkedException').
Revision ChangesPath
1.2 +33 -29jboss-j2ee/src/main/javax/jms/JMSException.java
Index: JMSException.java
===
RCS file: /cvsroot/jboss/jboss-j2ee/src/main/javax/jms/JMSException.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -u -r1.1 -r1.2
--- JMSException.java 2001/04/01 13:43:20 1.1
+++ JMSException.java 2001/08/21 23:47:10 1.2
@@ -10,24 +10,33 @@
/**
* This is the root class of all JMS exceptions.
*
- * It provides following information:
- *
+ * pIt provides following information:
* ul
- * liprovider-specific string describing the error - This string is the
standard Java exception message, and is available via getMessage().
+ * liprovider-specific string describing the error -
+ * This string is the standard Java exception message,
+ * and is available via getMessage().
* liprovider-specific, string error code
- * lireference to another exception - Often a JMS exception will be the result
of a lower level problem. If appropriate, this lower level exception can be linked to
the JMS exception.
+ * lireference to another exception - Often a JMS exception will be
+ * the result of a lower level problem. If appropriate, this lower
+ * level exception can be linked to the JMS exception.
* /ul
*
* @author Chris Kimpton ([EMAIL PROTECTED])
- * @version $Revision: 1.1 $
+ * @author a href=mailto:[EMAIL PROTECTED];Jason Dillon/a
+ * @version $Revision: 1.2 $
**/
-public class JMSException extends Exception
+public class JMSException
+ extends Exception
{
- // CONSTRUCTORS -
+ /** The specified error code */
+ private String _errorCode; // = null;
+ /** A linked exception */
+ private Exception _linkedException;
+
/**
- * Construct a JMSException with reason and error code for exception
- */
+* Construct a JMSException with reason and error code for exception
+*/
public JMSException(String reason, String errorCode)
{
super(reason);
@@ -35,40 +44,35 @@
}
/**
- * Construct a JMSException with reason and with error code defaulting to null
- */
+* Construct a JMSException with reason and with error code defaulting
+* to null.
+*/
public JMSException(String reason)
{
- this(reason,null);
+ this(reason, null);
}
-
- // PUBLIC METHODS -
- /** Get the vendor specific error code */
+ /**
+* Get the vendor specific error code.
+*/
public String getErrorCode()
{
return _errorCode;
}
- /** Get the exception linked to this one.
- * ? who uses this ?
- */
+ /**
+* Get the exception linked to this one.
+*/
public Exception getLinkedException()
{
return _linkedException;
}
- /** set the linked exception */
- public void setLinkedException(Exception linkedException)
+ /**
+* Set the linked exception
+*/
+ public synchronized void setLinkedException(Exception ex)
{
- _linkedException = linkedException;
+ _linkedException = ex;
}
-
- // ATTRIBUTES -
-
- /** the specified error code */
- private String _errorCode = null;
-
- /** a linked exception */
- private Exception _linkedException;
}
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jboss-j2ee/src/main/javax/jms JMSException.java
User: user57
Date: 01/08/21 16:51:22
Modified:src/main/javax/jms JMSException.java
Log:
o needed to change field names too (removed leading _)
Revision ChangesPath
1.3 +7 -7 jboss-j2ee/src/main/javax/jms/JMSException.java
Index: JMSException.java
===
RCS file: /cvsroot/jboss/jboss-j2ee/src/main/javax/jms/JMSException.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- JMSException.java 2001/08/21 23:47:10 1.2
+++ JMSException.java 2001/08/21 23:51:22 1.3
@@ -23,16 +23,16 @@
*
* @author Chris Kimpton ([EMAIL PROTECTED])
* @author a href=mailto:[EMAIL PROTECTED];Jason Dillon/a
- * @version $Revision: 1.2 $
+ * @version $Revision: 1.3 $
**/
public class JMSException
extends Exception
{
/** The specified error code */
- private String _errorCode; // = null;
+ private String errorCode; // = null;
/** A linked exception */
- private Exception _linkedException;
+ private Exception linkedException;
/**
* Construct a JMSException with reason and error code for exception
@@ -40,7 +40,7 @@
public JMSException(String reason, String errorCode)
{
super(reason);
- _errorCode = errorCode;
+ this.errorCode = errorCode;
}
/**
@@ -57,7 +57,7 @@
*/
public String getErrorCode()
{
- return _errorCode;
+ return errorCode;
}
/**
@@ -65,7 +65,7 @@
*/
public Exception getLinkedException()
{
- return _linkedException;
+ return linkedException;
}
/**
@@ -73,6 +73,6 @@
*/
public synchronized void setLinkedException(Exception ex)
{
- _linkedException = ex;
+ linkedException = ex;
}
}
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmq/src/main/org/jboss/mq SpyConnectionConsumer.java
User: dmaplesden
Date: 01/08/21 17:07:18
Modified:src/main/org/jboss/mq SpyConnectionConsumer.java
Log:
Implement the loading of multiple messages into each session when under heavy load.
Revision ChangesPath
1.5 +26 -9 jbossmq/src/main/org/jboss/mq/SpyConnectionConsumer.java
Index: SpyConnectionConsumer.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/SpyConnectionConsumer.java,v
retrieving revision 1.4
retrieving revision 1.5
diff -u -r1.4 -r1.5
--- SpyConnectionConsumer.java2001/08/17 03:04:01 1.4
+++ SpyConnectionConsumer.java2001/08/22 00:07:18 1.5
@@ -19,7 +19,7 @@
*
* @author Hiram Chirino ([EMAIL PROTECTED])
* @createdAugust 16, 2001
- * @version$Revision: 1.4 $
+ * @version$Revision: 1.5 $
*/
public class SpyConnectionConsumer implements javax.jms.ConnectionConsumer,
SpyConsumer, Runnable {
@@ -60,6 +60,8 @@
this.destination = destination;
this.serverSessionPool = serverSessionPool;
this.maxMessages = maxMessages;
+ if(this.maxMessages 1)
+ this.maxMessages = 1;
subscription.destination = ( SpyDestination )destination;
subscription.messageSelector = messageSelector;
@@ -142,13 +144,24 @@
//Used to facilitate delivery of messages to sessions from server session pool.
public void run() {
- SpyMessage mes = null;
+ java.util.ArrayList mesList = new java.util.ArrayList();
try {
outer :
while ( true ) {
-//get Message
-while ( mes == null ) {
- mes = connection.receive( subscription, 0 );
+synchronized( queue ){
+ if(closed)
+ break outer;
+}
+//get Messages
+for(int i=0;imaxMessages;i++){
+ SpyMessage mes = connection.receive(subscription, -1); //receive no
wait
+ if(mes == null)
+ break;
+ else
+ mesList.add(mes);
+}
+if(mesList.isEmpty()){
+ SpyMessage mes = connection.receive( subscription, 0 );
if ( mes == null ) {
synchronized ( queue ) {
waitingForMessage = true;
@@ -166,7 +179,9 @@
waitingForMessage = false;
}
}
+ mesList.add(mes);
}
+
ServerSession serverSession = serverSessionPool.getServerSession();
SpySession spySession = ( SpySession )serverSession.getSession();
@@ -176,17 +191,19 @@
spySession.sessionConsumer.subscription = subscription;
}
-spySession.addMessage( mes );
+for(int i=0;imesList.size();i++){
+ spySession.addMessage( (SpyMessage)mesList.get(i) );
+}
cat.debug( + this + Starting the ServerSession. );
serverSession.start();
-mes = null;
+mesList.clear();
}
} catch ( JMSException e ) {
cat.warn( Connection consumer closing due to error in listening thread.,
e );
try {
-if ( mes != null ) {
- connection.send( mes.getAcknowledgementRequest( false ) );
+for(int i=0;imesList.size();i++){
+ connection.send(
((SpyMessage)mesList.get(i)).getAcknowledgementRequest( false ) );
}
close();
} catch ( Exception ignore ) {
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] manual build
Interesting, today I worked on the manual build also. I have it working with jdk 1.3.1, ant 1.3, some updated xml jars. problems I still have: html files end up in manual instead of manual/output/html jbossjms.xml has a syntax error. Problems I fixed: build.sh does not set ANT_HOME correctly, so the script uses a system ant instead of the one in jboss. XSLTransform, in ant.jar, uses several classes from Optional.jar, so build.sh also has to include optional.jar on the local classpath. The xml processor files in tools... are out of date versions. I hope to get the output directory fixed shortly, probably the syntax error will be easy too. Since mine works with ant 1.3 should I commit? or should we be forward looking? david jencks On 2001.08.21 18:56:25 -0400 Vincent Harcq wrote: Hi, Is there anybody trying to make the manual build run correctly? I give it a (painful) try. I succeed to build a part of html and printable-html (hum simply the preface,intro, cmp and jaws for the moment) I choose to move to ant 1.4 that use JAXP 1.1 to avoid problems with style, xsl, ... I needed to change the jboss-all/build/build.bat to include sun/jaxp and ant/lib/ext jars. The manual build.xml has also changed. In jbossdocs.xml I have to reference the _entity_ using full path Because of these (ant 1.4 is beta and I did not check compliance with other modules), I don't know if it is a good idea to commit now. What do you think ? Create a branch until it works ? The problems I have now: printable-html : Some javax.xml.transform.TransformerException when I added more chapter xml files to jbossdocs.xml. printable-html : the CSS is nt included in generated html. printable-html : the file is jbossdocshtml (the . is missing) html: it fails everytime with a javax.xml.transform.TransformerException: java.lang.ArrayIndexOutOfBoundsException at the end of the build. Vincent. _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] jbossmq load balancing across vms
I am not an expert Is this sharing of a message queue receiving end really spec compliant? How does it relate to order-of-messages guarantees? It seems to me this may be the conceptual difference between message queues and javaspaces..?? david jencks On 2001.08.21 19:06:08 -0400 Jason Dillon wrote: Hello again, I am running into a problem with the latest JBossMQ wrt using it to load balance across machines. I have one machine which acts as a message router, running JBossMQ. All other machines use it for their JMSProvider. When I have more than one MDB attached to a single queue, and a user-thread adding messages over and over, it looks like only one of the MDB's will ever read any messages. I do not remember seeing this before. It looks like when I startup a new MDB that 50% of the time it will take over and the others will remain idle, or it will simply never read any new messages. I have each of these MDB setup with: MaximumSize1/MaximumSize MaxMessages1/MaxMessages Which could be causing problem... but honestly I don't really know what these two setting actually do. I wanted to have these deployed MDB's to only process one message at a time, which this seems to accomplish most of the time. Any ways, I am concerended about how JBossMQ is pushing messages over to consumers. I mentioned before that if one producer added 1k messages to a queue, then one MDB was attached which would suck up all of the messages. When a second MDB was attached it would not have anything todo. Hiram mentioned something about this, but I am not really sure what the plans were to fix this. I would like to get this fixed soon, as I am fairly dependent on this behavior. From my perspective, there are two major issues: 1) JBossMQ does not seem to trade off between consumers (ie. round-robin or whatever). 2) JBossMQ does not have any configurable limits on how messages are pushed from the server to clients/consumers. I understand why JBossMQ wants to push over bulk messages to consumers, but that is only really useful in some applications. In my case, I would rather that it pushed over smallers chunks (or even 1 msg at a time) to clients so that I could utilize the cluster of machines that I have to distribute the processing load. So, to the JBossMQ folks, does this sound like it would be easy to implement? Do you have any suggestions on how it might be done (quick fix and full blown god JMS provider version)? I am willing to look into this more, but I could use some pointers on which bits need to be tweaked. Any ideas? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] jbossmq selector parser grammer source
Where is the source for the selector parser? It looks like it is not parsing the literals true and false, it only looks for TRUE and FALSE. I am contemplating modifiying parser.java int yylex() under the //CST group group, but I would rather modify the grammer source and regenerate the parser. Perhaps we could switch this to javacc and let ant build the parser as part of the build. It looks like there might be a problem with Operator.equal() too, which I am looking into now... Yup, sure is. Operator.equal() does not properly handle Boolean types. I have just fixed (and verified) both of these problems. Should I commit them? Silly me, of course I should commit them... but where is jms.y? When ever the parser is regenerated from this it will blow away these changes. Can someone change this to javacc, check in the source grammer and setup the build system to automatically generate this? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmq/src/main/org/jboss/mq/selectors Operator.java
User: user57
Date: 01/08/21 17:44:44
Modified:src/main/org/jboss/mq/selectors Operator.java
Log:
o Operator.equal() will properly compute the equality when using Boolean
types.
Revision ChangesPath
1.3 +3 -1 jbossmq/src/main/org/jboss/mq/selectors/Operator.java
Index: Operator.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/selectors/Operator.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- Operator.java 2001/08/17 03:04:06 1.2
+++ Operator.java 2001/08/22 00:44:44 1.3
@@ -14,7 +14,7 @@
* @author Norbert Lataille ([EMAIL PROTECTED])
* @author [EMAIL PROTECTED]
* @createdAugust 16, 2001
- * @version$Revision: 1.2 $
+ * @version$Revision: 1.3 $
*/
public class Operator {
int operation;
@@ -183,7 +183,9 @@
case STRING:
case LONG:
case DOUBLE:
+ case BOOLEAN:
computeArgument2();
+
if ( arg2 == null ) {
return Boolean.FALSE;
}
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmq/src/main/org/jboss/mq/selectors parser.java
User: user57
Date: 01/08/21 17:52:31
Modified:src/main/org/jboss/mq/selectors parser.java
Log:
o Added support for the true and false boolean literals.
Revision ChangesPath
1.3 +2 -2 jbossmq/src/main/org/jboss/mq/selectors/parser.java
Index: parser.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/selectors/parser.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- parser.java 2001/08/17 03:04:06 1.2
+++ parser.java 2001/08/22 00:52:31 1.3
@@ -540,11 +540,11 @@
}
//CST group
- if ( s.equals( TRUE ) ) {
+ if ( s.equals( TRUE ) || s.equals(true) ) {
yylval = new parserval( ( Object )Boolean.TRUE );
return CST;
}
- if ( s.equals( FALSE ) ) {
+ if ( s.equals( FALSE ) || s.equals(false) ) {
yylval = new parserval( ( Object )Boolean.FALSE );
return CST;
}
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] CVS update: jbossmq/src/main/org/jboss/mq/selectors Selector.java
User: user57
Date: 01/08/21 17:53:35
Modified:src/main/org/jboss/mq/selectors Selector.java
Log:
o added some commented logging (too verbose to leave it for now)
o changed exceptions to return a bit more detail
Revision ChangesPath
1.3 +105 -62 jbossmq/src/main/org/jboss/mq/selectors/Selector.java
Index: Selector.java
===
RCS file: /cvsroot/jboss/jbossmq/src/main/org/jboss/mq/selectors/Selector.java,v
retrieving revision 1.2
retrieving revision 1.3
diff -u -r1.2 -r1.3
--- Selector.java 2001/08/17 03:04:06 1.2
+++ Selector.java 2001/08/22 00:53:35 1.3
@@ -8,10 +8,12 @@
import java.util.HashMap;
import java.util.Iterator;
-import javax.jms.InvalidSelectorException;
+import javax.jms.InvalidSelectorException;
import javax.jms.JMSException;
+import org.apache.log4j.Category;
+
import org.jboss.mq.SpyMessage;
/**
@@ -19,112 +21,153 @@
*
* @author Norbert Lataille ([EMAIL PROTECTED])
* @author Juha Lindfors ([EMAIL PROTECTED])
+ * @author a href=mailto:[EMAIL PROTECTED];Jason Dillon/a
* @createdAugust 16, 2001
- * @version$Revision: 1.2 $
+ * @version$Revision: 1.3 $
*/
-public class Selector {
- public HashMap identifiers;
- // HashMap of Identifiers
- public Objectresult;
+public class Selector
+{
+ public HashMap identifiers;
+ public Object result;
+ private Class resultType;
+
+ static Category cat = Category.getInstance(Selector.class);
- static org.apache.log4j.Category cat = org.apache.log4j.Category.getInstance(
Selector.class );
-
- public Selector( String sel )
- throws JMSException {
+ public Selector(String sel) throws JMSException {
parser bob = new parser();
identifiers = new HashMap();
try {
- result = bob.parse( sel, identifiers );
- } catch ( Exception e ) {
- InvalidSelectorException exception = new InvalidSelectorException( The
selector is invalid. );
- exception.setLinkedException( e );
+ result = bob.parse(sel, identifiers);
+ resultType = result.getClass();
+ }
+ catch (Exception e) {
+ InvalidSelectorException exception =
+new InvalidSelectorException(The selector is invalid.);
+ exception.setLinkedException(e);
throw exception;
}
- //Log.notice(result.toString());
+// if (cat.isDebugEnabled()) {
+// cat.debug(result: + resultType + = + result);
+// }
}
-
- public boolean test( SpyMessage mes )
- throws JMSException {
+ public boolean test(SpyMessage mes) throws JMSException {
try {
-
- //Set the identifiers values
+ // Set the identifiers values
Iterator i = identifiers.values().iterator();
-
- while ( i.hasNext() ) {
-Identifier id = ( Identifier )i.next();
-Object find = mes.getObjectProperty( id.name );
+ while (i.hasNext()) {
+Identifier id = (Identifier)i.next();
+Object find = mes.getObjectProperty(id.name);
+// if (cat.isDebugEnabled()) {
+//cat.debug(Identifier: + id);
+//cat.debug(Property: + find);
+//if (find != null) {
+// cat.debug(Property type: + find.getClass());
+//}
+// }
-if ( find == null ) {
- find = getHeaderFieldReferences( mes, id.name );
+if (find == null) {
+ find = getHeaderFieldReferences(mes, id.name);
}
-if ( find == null ) {
- cat.debug( Warning : missing property + id.name );
+if (find == null) {
+ if (cat.isDebugEnabled()) {
+ cat.debug(Warning : missing property + id.name);
+ }
id.value = null;
-} else {
+}
+else {
Class type = find.getClass();
- if ( type.equals( Boolean.class ) || type.equals( String.class )
- || type.equals( Double.class ) || type.equals( Float.class )
- || type.equals( Integer.class ) || type.equals( Long.class )
- || type.equals( Short.class ) || type.equals( Byte.class ) ) {
+ if (type.equals(Boolean.class) ||
+ type.equals(String.class) ||
+ type.equals(Double.class) ||
+ type.equals(Float.class) ||
+ type.equals(Integer.class) ||
+ type.equals(Long.class)||
+ type.equals(Short.class)
Re: [JBoss-dev] manual build
What xml jars did you have to update? Where did you update them? XSLTransform, in ant.jar, uses several classes from Optional.jar, so build.sh also has to include optional.jar on the local classpath. By adding optional.jar to the system classpath, all support jars will also have to be on the local classpath, which means that users will either have to use a custom ant, which contains all the correct bits in $ANT_HOME/lib or they will have to install the required versions of those jars into their $ANT_HOME/lib or the build system will not work. This is a rather huge problem with Ant 1.3 and how it loads classes. The easiest thing todo is insist that all users use the tools/apache/ant/bin/ant* to start the build (via build.* scripts) or they can update there own ANT_HOME to include all of the jars from tools/apache/ant/lib into their $ANT_HOME/lib. I origionally had things setup to use a version of ant with the correct libs, but reverted to the current style to allow people to run any standard ant dist directly. Unfortunatly the optional.jar makes this whole thing a huge mess. I do not have a good fix for this at the moment. Does anyone know if 1.4 fixes this crappy class loading problem? The xml processor files in tools... are out of date versions. True, but they are from the ant 1.3 dist, so if we expect users to run a vanilla ant and get it to work, we have to use these. Since mine works with ant 1.3 should I commit? or should we be forward looking? I think we need to get an answer about how 1.4 might fix this and allow users to still use `ant` directly, or if that requirement is causing more problems than it solves. --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] jbossmq load balancing across vms
It does not say... this is one of the grey-areas of the spec. Most providers will round-robin over multipule recievers. It is really too bad that this is not more concreate, as JMS would provide a nice mech. for distributing over a large group of machines, and allow for ordering and such. I think that I will probably drop JMS for this as soon as I have time to rearch. under a different distribution system (like JavaSpaces). JMS is too ugly and too vendor specific. I will probably still keep it around for async events from beans/to beans though. Just not as the backend for distributing load. --jason On Tue, 21 Aug 2001, David Jencks wrote: I am not an expert Is this sharing of a message queue receiving end really spec compliant? How does it relate to order-of-messages guarantees? It seems to me this may be the conceptual difference between message queues and javaspaces..?? david jencks On 2001.08.21 19:06:08 -0400 Jason Dillon wrote: Hello again, I am running into a problem with the latest JBossMQ wrt using it to load balance across machines. I have one machine which acts as a message router, running JBossMQ. All other machines use it for their JMSProvider. When I have more than one MDB attached to a single queue, and a user-thread adding messages over and over, it looks like only one of the MDB's will ever read any messages. I do not remember seeing this before. It looks like when I startup a new MDB that 50% of the time it will take over and the others will remain idle, or it will simply never read any new messages. I have each of these MDB setup with: MaximumSize1/MaximumSize MaxMessages1/MaxMessages Which could be causing problem... but honestly I don't really know what these two setting actually do. I wanted to have these deployed MDB's to only process one message at a time, which this seems to accomplish most of the time. Any ways, I am concerended about how JBossMQ is pushing messages over to consumers. I mentioned before that if one producer added 1k messages to a queue, then one MDB was attached which would suck up all of the messages. When a second MDB was attached it would not have anything todo. Hiram mentioned something about this, but I am not really sure what the plans were to fix this. I would like to get this fixed soon, as I am fairly dependent on this behavior. From my perspective, there are two major issues: 1) JBossMQ does not seem to trade off between consumers (ie. round-robin or whatever). 2) JBossMQ does not have any configurable limits on how messages are pushed from the server to clients/consumers. I understand why JBossMQ wants to push over bulk messages to consumers, but that is only really useful in some applications. In my case, I would rather that it pushed over smallers chunks (or even 1 msg at a time) to clients so that I could utilize the cluster of machines that I have to distribute the processing load. So, to the JBossMQ folks, does this sound like it would be easy to implement? Do you have any suggestions on how it might be done (quick fix and full blown god JMS provider version)? I am willing to look into this more, but I could use some pointers on which bits need to be tweaked. Any ideas? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] jbossmq load balancing across vms
I'm not up-to-date on javaspaces but with jms you can certainly have multiple receivers on the same message queue, you can also have multiple senders. The JMS spec does not specify any policy for how messages are distributed between different receivers though. As for message order, it is only guaranteed amongst messages sent to the same queue from the same sender, that is if you receive two messages from the same queue sent by the same sender they will arrive in the order sent (except for some cases involving different priorities and the like). Cheers, David -Original Message- From: David Jencks [mailto:[EMAIL PROTECTED]] Sent: Wednesday, August 22, 2001 12:48 PM To: [EMAIL PROTECTED] Subject: Re: [JBoss-dev] jbossmq load balancing across vms I am not an expert Is this sharing of a message queue receiving end really spec compliant? How does it relate to order-of-messages guarantees? It seems to me this may be the conceptual difference between message queues and javaspaces..?? david jencks On 2001.08.21 19:06:08 -0400 Jason Dillon wrote: Hello again, I am running into a problem with the latest JBossMQ wrt using it to load balance across machines. I have one machine which acts as a message router, running JBossMQ. All other machines use it for their JMSProvider. When I have more than one MDB attached to a single queue, and a user-thread adding messages over and over, it looks like only one of the MDB's will ever read any messages. I do not remember seeing this before. It looks like when I startup a new MDB that 50% of the time it will take over and the others will remain idle, or it will simply never read any new messages. I have each of these MDB setup with: MaximumSize1/MaximumSize MaxMessages1/MaxMessages Which could be causing problem... but honestly I don't really know what these two setting actually do. I wanted to have these deployed MDB's to only process one message at a time, which this seems to accomplish most of the time. Any ways, I am concerended about how JBossMQ is pushing messages over to consumers. I mentioned before that if one producer added 1k messages to a queue, then one MDB was attached which would suck up all of the messages. When a second MDB was attached it would not have anything todo. Hiram mentioned something about this, but I am not really sure what the plans were to fix this. I would like to get this fixed soon, as I am fairly dependent on this behavior. From my perspective, there are two major issues: 1) JBossMQ does not seem to trade off between consumers (ie. round-robin or whatever). 2) JBossMQ does not have any configurable limits on how messages are pushed from the server to clients/consumers. I understand why JBossMQ wants to push over bulk messages to consumers, but that is only really useful in some applications. In my case, I would rather that it pushed over smallers chunks (or even 1 msg at a time) to clients so that I could utilize the cluster of machines that I have to distribute the processing load. So, to the JBossMQ folks, does this sound like it would be easy to implement? Do you have any suggestions on how it might be done (quick fix and full blown god JMS provider version)? I am willing to look into this more, but I could use some pointers on which bits need to be tweaked. Any ideas? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
[JBoss-dev] [ jboss-Change Notes-454045 ] JBossMQ message selector fix
Change Notes item #454045, was opened at 2001-08-21 18:24 You can respond by visiting: http://sourceforge.net/tracker/?func=detailatid=381174aid=454045group_id=22866 Category: None Group: None Status: Open Priority: 5 Submitted By: Jason Dillon (user57) Assigned to: Nobody/Anonymous (nobody) Summary: JBossMQ message selector fix Initial Comment: The JBossMQ message selector parser will now allow the boolean literals true and false as well as TRUE and FALSE. Equality operators can now be used on Boolean properties values too. For example, if a boolean property is set on a message: msg.setBooleanProperty(myProperty, true); The following message selector will function correctly: myProperty = true or myProperty = TRUE -- You can respond by visiting: http://sourceforge.net/tracker/?func=detailatid=381174aid=454045group_id=22866 ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] jbossmq load balancing across vms
What this means is that if you have multiple receivers waiting for a message from a queue and a message arrives then it will be immediately delivered to receiver A. Receiver A is now removed from the HashSet, however if receiver A finishes its processing and requests another message before another message is sent to the queue it will be added back to the HashSet. When the next message arrives set will again be sent to receiver A because in the HashSets ordering receiver A is clearly ahead of receiver B. Why a HashSet and not a LinkedList? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
Re: [JBoss-dev] manual build
On 2001.08.21 21:07:53 -0400 Jason Dillon wrote:
What xml jars did you have to update? Where did you update them?
I'll get back to you on that- mostly remove parser.jar and include
xalan/xerces from previous manual build.
XSLTransform, in ant.jar, uses several classes from Optional.jar, so
build.sh also has to include optional.jar on the local classpath.
By adding optional.jar to the system classpath, all support jars will
also
have to be on the local classpath, which means that users will either
have
to use a custom ant, which contains all the correct bits in $ANT_HOME/lib
or
they will have to install the required versions of those jars into their
$ANT_HOME/lib or the build system will not work.
Well, after making search set ANT_HOME instead of ANT, so the result is
useful, I put this at the end of build.sh and the manual builds more or
less
echo ant_home is $ANT_HOME
#kill preexisting classpath, we want our just found ant.
CLASSPATH=
#for manual only, include optional.jar on classpath- lousy ant style
task
#design has XSLProcess in ant.jar and loaded classes in optional.jar!!!
CLASSPATH=$ANT_HOME/lib/ext/optional.jar
export ANT ANT_HOME CLASSPATH
exec $ANT $ANT_OPTIONS $@
}
##
## Bootstrap
##
main $@
the ant script takes care of adding ant.jar to the classpath.
I'm a little confused. Are you intentionally including an ant distro in
jboss yet having the build.sh scripts not find it? It seems to me that the
jboss build system, since it includes a copy of ant, should use that copy
by default, and find it properly.
david jencks
ps.
here's my change to search:
search() {
search=$*
for d in $search; do
ANT_HOME=`pwd`/$d
ANT=$ANT_HOME/bin/ant
if [ -x $ANT ]; then
# found one
#this is return value of function!
echo $ANT_HOME
break
fi
done
}
Returning ANT_HOME lets ANT_HOME be set, and ANT can be defined from it.
Returning ANT, as in cvs, seems to be useless -- ANT_HOME is never set, so
the ANT just set can't be found. At any rate this seemed to be what was
happening on my system.
This is a rather huge problem with Ant 1.3 and how it loads classes. The
easiest thing todo is insist that all users use the
tools/apache/ant/bin/ant* to start the build (via build.* scripts) or
they
can update there own ANT_HOME to include all of the jars from
tools/apache/ant/lib into their $ANT_HOME/lib.
I origionally had things setup to use a version of ant with the correct
libs, but reverted to the current style to allow people to run any
standard
ant dist directly.
Unfortunatly the optional.jar makes this whole thing a huge mess. I do
not
have a good fix for this at the moment. Does anyone know if 1.4 fixes
this
crappy class loading problem?
The xml processor files in tools... are out of date versions.
True, but they are from the ant 1.3 dist, so if we expect users to run a
vanilla ant and get it to work, we have to use these.
Since mine works with ant 1.3 should I commit? or should we be forward
looking?
I think we need to get an answer about how 1.4 might fix this and allow
users to still use `ant` directly, or if that requirement is causing more
problems than it solves.
--jason
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
___
Jboss-development mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] jbossmq load balancing across vms
Basically because of efficiency. Because the possibility exists that when a receive is done the receiver may already be in the list of receivers (due to receives with timeouts) we have to check that a receiver is not already in the list before adding them to the list. This search O(n) for a LinkedList and constant time for a HashSet. If you are absolutely desparate for the round-robin allocation of messages then I would re-implement keeping both the LinkedList and the HashSet. Use the HashSet to check if the receiver is already receiving and the LinkedList to maintain the order of the requests. The code I am talking about is the org.jboss.mq.server.BasicQueue class. Cheers, David. -Original Message- From: Jason Dillon [mailto:[EMAIL PROTECTED]] Sent: Wednesday, August 22, 2001 1:50 PM To: '[EMAIL PROTECTED]' Subject: RE: [JBoss-dev] jbossmq load balancing across vms What this means is that if you have multiple receivers waiting for a message from a queue and a message arrives then it will be immediately delivered to receiver A. Receiver A is now removed from the HashSet, however if receiver A finishes its processing and requests another message before another message is sent to the queue it will be added back to the HashSet. When the next message arrives set will again be sent to receiver A because in the HashSets ordering receiver A is clearly ahead of receiver B. Why a HashSet and not a LinkedList? --jason ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
RE: [JBoss-dev] manual build
Hi, Interesting, today I worked on the manual build also. I have it working with jdk 1.3.1, ant 1.3, some updated xml jars. problems I still have: html files end up in manual instead of manual/output/html I had to change the copy todir task and some others as well. If you run one by one (html only or printablehtml only), it simpler to see where the html are generated. jbossjms.xml has a syntax error. Problems I fixed: build.sh does not set ANT_HOME correctly, so the script uses a system ant instead of the one in jboss. XSLTransform, in ant.jar, uses several classes from Optional.jar, so build.sh also has to include optional.jar on the local classpath. The xml processor files in tools... are out of date versions. That's the reason why I switched to 1.4. My build (.bat) is doing : for %%i in (..\thirdparty\sun\jaxp\lib\*.jar) do call lcp.bat %%i for %%i in (..\tools\apache\ant\lib\*.jar) do call lcp.bat %%i for %%i in (..\tools\apache\ant\lib\ext\*.jar) do call lcp.bat %%i With no more xml jars in ant\lib\ I hope to get the output directory fixed shortly, probably the syntax error will be easy too. Since mine works with ant 1.3 should I commit? or should we be forward looking? david jencks On 2001.08.21 18:56:25 -0400 Vincent Harcq wrote: Hi, Is there anybody trying to make the manual build run correctly? I give it a (painful) try. I succeed to build a part of html and printable-html (hum simply the preface,intro, cmp and jaws for the moment) I choose to move to ant 1.4 that use JAXP 1.1 to avoid problems with style, xsl, ... I needed to change the jboss-all/build/build.bat to include sun/jaxp and ant/lib/ext jars. The manual build.xml has also changed. In jbossdocs.xml I have to reference the _entity_ using full path Because of these (ant 1.4 is beta and I did not check compliance with other modules), I don't know if it is a good idea to commit now. What do you think ? Create a branch until it works ? The problems I have now: printable-html : Some javax.xml.transform.TransformerException when I added more chapter xml files to jbossdocs.xml. printable-html : the CSS is nt included in generated html. printable-html : the file is jbossdocshtml (the . is missing) html: it fails everytime with a javax.xml.transform.TransformerException: java.lang.ArrayIndexOutOfBoundsException at the end of the build. Vincent. _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development _ Do You Yahoo!? Get your free @yahoo.com address at http://mail.yahoo.com ___ Jboss-development mailing list [EMAIL PROTECTED] http://lists.sourceforge.net/lists/listinfo/jboss-development
