Re[2]: [JBoss-user] CMR Error under load

[Alex L. <lou...@ua...>]

with fresh 3.2.0RC5 from CVS?

alex

Tuesday, April 01, 2003, 11:26:47 AM, Simone Milani wrote:

SM> Hi Andrew,

SM>     I had the error as well....


SM> Simone


SM> ----- Original Message -----
SM> From: "Andrew May" <am...@in...>
SM> To: <jbo...@li...>
SM> Sent: Tuesday, April 01, 2003 8:40 AM
SM> Subject: Re: [JBoss-user] CMR Error under load


>> Simone,
>>
>> I manually inserted a couple of records. e.g.:
>>
>> INSERT INTO Parent VALUES ('test1');
>> INSERT INTO Child (id, name, value, parentid) VALUES ('test1:name1',
SM> 'name1', 'value1',
>> 'test1');
>>
>> And that was enough data to recreate the problem. The test was then called
SM> with
>>
>> java com.ingenta.acs.test.ejb.CMRTest test1 30 100
>>
>> - that's with a parent pk of test1, 30 threads each doing 100 loops. The
SM> classpath has
>> test.jar, all the necessary JBOSS client jars, and a jndi.properties
SM> containing a provider
>> url.
>>
>> For every successful record it will print the returned name/value pairs
SM> from the child.
>> For every failure it will print the error message (the full stack trace
SM> should be in the
>> server log).
>>
>> Thanks for having a look at this - I've had to wait for our firewall to be
SM> opened up
>> before I can build the latest 3.2 code from CVS.
>>
>> Andrew
>>
>> Simone Milani wrote:
>>
>> > Hi Andrew,
>> >
>> >     I am trying your test, do you have a script to populate the tables
SM> and
>> > the parameters for the CMRTestClass?
>> >
>> > Thanks
>> >
>> > Simone
>> > ----- Original Message -----
>> > From: "Andrew May" <am...@in...>
>> > To: <jbo...@li...>
>> > Sent: Monday, March 31, 2003 2:46 PM
>> > Subject: Re: [JBoss-user] CMR Error under load
>> >
>> >
>> >
>> >>I don't see a RC5 release on sourceforge, so I assume this means I have
SM> to
>> >
>> > build it from
>> >
>> >>CVS. I'll give that a go, but as I've never tried it before it could
SM> take
>> >
>> > me a while
>> >
>> >>(never actually used CVS I confess) - still it's something I ought to be
>> >
>> > able to do.
>> >
>> >>If anyone has a chance to run my test on this version of JBoss I'd be
>> >
>> > grateful - it'll
>> >
>> >>probably be at least a day before I get that far.
>> >>
>> >>Andrew
>> >>
>> >>Alex Loubyansky wrote:
>> >>
>> >>
>> >>>Could you, please, try it on JBoss-3.2.0RC5 which is the current
>> >>>version.
>> >>>I have a testcase that failed in my old 3.2 and passes in current
>> >>>3.2.0RC5.
>> >>>
>> >>>Please, try it.
>> >>>
>> >>>alex
>> >>>
>> >>>Monday, March 31, 2003, 12:37:32 PM, Andrew May wrote:
>> >>>
>> >>>AM> OK, here's a test case that recreates the problem for me.
>> >>>
>> >>>AM> I've attached test.jar which is the ejb-jar and also includes the
>> >
>> > test case class. Also
>> >
>> >>>AM> attached is test-src.jar which has the source code for the EJBs and
>> >
>> > the test. I'm afraid
>> >
>> >>>AM> I've been hacking around existing build scripts to create the test
>> >
>> > so I've not included
>> >
>> >>>AM> any way of building the code. I'm using an older version of
SM> XDoclet.
>> >>>
>> >>>AM> There are 2 entities, Parent and Child. The CMRTest class creates
>> >
>> > multiple threads to call
>> >
>> >>>AM> the getPropertyMap() method on the Parent. This method iterates
>> >
>> > around the CMR collection
>> >
>> >>>AM> of children, and when multiple threads are run causes occasional
SM> CMR
>> >
>> > errors.
>> >
>> >>>AM> In jbosscmp-jdbc.xml the datasource is java:/jdbc.testPool and the
>> >
>> > mapping is for Oracle8.
>> >
>> >>>AM> My test doesn't create the test data, so I've manually inserted
SM> data
>> >
>> > before running the
>> >
>> >>>AM> test. I've found that it's only necessary to have one child record
>> >
>> > to recreate the problem.
>> >
>> >>>AM> The CMRTest class takes 3 arguments - the parent primary key,
SM> number
>> >
>> > of threads, number of
>> >
>> >>>AM> loops per thread. e.g.:
>> >>>
>> >>>AM> java com.ingenta.acs.test.ejb.CMRTest test1 30 100
>> >>>
>> >>>AM> 30 threads with 100 loops always causes at least one CMR error for
>> >
>> > me. I've tested on two
>> >
>> >>>AM> setups, both of which cause errors.
>> >>>
>> >>>AM> 1) Solaris, Sun JDK1.3.1, JBoss 3.2.0RC3, Oracle 8.1.7.2.0
>> >>>AM> 2) Linux (Mandrake 9), Sun JDK1.4.1, JBoss 3.2.0RC4, Oracle
>> >
>> > 8.1.7.2.0
>> >
>> >>>AM> jboss.xml in the ejb-jar declares all get* methods to be read-only.
>> >
>> > If I remove this the
>> >
>> >>>AM> CMR error does not occur, but the test goes a lot slower.
>> >>>
>> >>>AM> Let me know if there's any more information you need.
>> >>>
>> >>>AM> Andrew
>> >>>
>> >>>AM> Dain Sundstrom wrote:
>> >>>
>> >>>
>> >>>
>> >>>>>I fixed a similar bug in this stuff for Sacha on March 19th (version
>> >>>>>1.43.2.13).  The problem was registering for transaction
>> >
>> > synchronization
>> >
>> >>>>>on the transaction in the context and not the transaction associated
>> >>>>>with the current thread.  This bug created the type of problem you
SM> have
>> >>>>>described, but the fix should bin in 3.2 rc4.
>> >>>>>
>> >>>>>Until someone can create a test case that reproduces the new bug, I
>> >>>>>don't think there is much we can do.
>> >>>>>
>> >>>>>-dain
>> >>>>>
>> >>>>>On Friday, March 28, 2003, at 10:01 AM, Alex Loubyansky wrote:
>> >>>>>
>> >>>>>
>> >>>>>
>> >>>>>>Hello Andrew,
>> >>>>>>
>> >>>>>>you are not alone complaining about this. But nobody could produce a
>> >>>>>>testcase that fail at least eventually.
>> >>>>>>
>> >>>>>>I am trying to reproduce it but can't for the moment.
>> >>>>>>
>> >>>>>>I have a Parent and 1000 of related Child entities.
>> >>>>>>I have 100 concurrent threads finding parent by primary key, getting
>> >>>>>>children and iterating through relation collection.
>> >>>>>>I use standard container for both Parent and Child.
>> >>>>>>In jboss.xml get methods are read-only for both Parent and Child.
>> >>>>>>
>> >>>>>>Any more info you can provide? Again a testcase would be great. Let
SM> it
>> >>>>>>fail not each time but eventually.
>> >>>>>>
>> >>>>>>Thank you,
>> >>>>>>alex