Re: [OJB-developers] Re: OJB in C/S mode: TransactionInProgressEx
Brought to you by:
thma
Menu
▾
▴
Re: [OJB-developers] Re: OJB in C/S mode: TransactionInProgressEx
|
From: <ri...@ya...> - 2002-02-01 22:17:06
|
Hello,
I have tried the C/S patch, and I am getting this exception when two =
users try to login at the same time:
-------------------------------------------------------------
[ODMG] ERROR: null
org.odmg.TransactionNotInProgressException
at =
ojb.broker.server.PersistenceBrokerClient.commitTransaction(Persisten
ceBrokerClient.java:192)
at ojb.odmg.oql.OQLQueryImpl.execute(OQLQueryImpl.java:230)
at =
charlie.ojb.ProyectoManagerOJB.getEmpleado(ProyectoManagerOJB.java)
at charlie.web.LogonAction.perform(LogonAction.java)
at =
org.apache.struts.action.ActionServlet.processActionPerform(ActionSer
vlet.java:1787)
at =
org.apache.struts.action.ActionServlet.process(ActionServlet.java:158
6)
at =
org.apache.struts.action.ActionServlet.doPost(ActionServlet.java:510)
at javax.servlet.http.HttpServlet.service(HttpServlet.java:760)
at javax.servlet.http.HttpServlet.service(HttpServlet.java:853)
at =
org.apache.catalina.core.ApplicationFilterChain.internalDoFilter(Appl
icationFilterChain.java:247)
at =
org.apache.catalina.core.ApplicationFilterChain.doFilter(ApplicationF
ilterChain.java:193)
at =
org.apache.catalina.core.StandardWrapperValve.invoke(StandardWrapperV
alve.java:243)
at =
org.apache.catalina.core.StandardPipeline.invokeNext(StandardPipeline
.java:566)
at =
org.apache.catalina.core.StandardPipeline.invoke(StandardPipeline.jav
a:472)
at =
org.apache.catalina.core.ContainerBase.invoke(ContainerBase.java:943)
at =
org.apache.catalina.core.StandardContextValve.invoke(StandardContextV
alve.java:201)
at =
org.apache.catalina.core.StandardPipeline.invokeNext(StandardPipeline
.java:566)
at =
org.apache.catalina.core.StandardPipeline.invoke(StandardPipeline.jav
a:472)
at =
org.apache.catalina.core.ContainerBase.invoke(ContainerBase.java:943)
at =
org.apache.catalina.core.StandardContext.invoke(StandardContext.java:
2344)
at =
org.apache.catalina.core.StandardHostValve.invoke(StandardHostValve.j
ava:164)
at =
org.apache.catalina.core.StandardPipeline.invokeNext(StandardPipeline
.java:566)
at =
org.apache.catalina.valves.ErrorDispatcherValve.invoke(ErrorDispatche
rValve.java:170)
at =
org.apache.catalina.core.StandardPipeline.invokeNext(StandardPipeline
.java:564)
at =
org.apache.catalina.valves.ErrorReportValve.invoke(ErrorReportValve.j
ava:170)
at =
org.apache.catalina.core.StandardPipeline.invokeNext(StandardPipeline
.java:564)
at =
org.apache.catalina.valves.AccessLogValve.invoke(AccessLogValve.java:
462)
at =
org.apache.catalina.core.StandardPipeline.invokeNext(StandardPipeline
.java:564)
at =
org.apache.catalina.core.StandardPipeline.invoke(StandardPipeline.jav
a:472)
at =
org.apache.catalina.core.ContainerBase.invoke(ContainerBase.java:943)
at =
org.apache.catalina.core.StandardEngineValve.invoke(StandardEngineVal
ve.java:163)
at =
org.apache.catalina.core.StandardPipeline.invokeNext(StandardPipeline
.java:566)
at =
org.apache.catalina.core.StandardPipeline.invoke(StandardPipeline.jav
a:472)
at =
org.apache.catalina.core.ContainerBase.invoke(ContainerBase.java:943)
at =
org.apache.catalina.connector.http.HttpProcessor.process(HttpProcesso
r.java:1011)
at =
org.apache.catalina.connector.http.HttpProcessor.run(HttpProcessor.ja
va:1106)
at java.lang.Thread.run(Thread.java:484)
-------------------------------------------------------------------------=
One of the users logs in, the other gets a stack trace.
I'm sure is not the singleton trying to initialize the omdg variable =
twice, because I'm using the "double check idiom".
These are the messages in the OJB C/S console:
[ojb.broker.server.RequestProcessor] ERROR: Client tries to open tx =
twice
[ojb.broker.server.RequestProcessor] ERROR: =
ojb.broker.server.Request@e35d5
[ojb.broker.server.RequestProcessor] ERROR: null
[ojb.broker.server.RequestProcessor] ERROR: null
[ojb.broker.server.RequestProcessor] ERROR: Client tries to open tx =
twice
[ojb.broker.server.RequestProcessor] ERROR: Client tries to open tx =
twice
Thanks for the Struts + OJB sample code. It's very interesting how he =
passes values between Actions and OJB, using a Hashtable.
Please let me know what to do next...
Thanks A LOT for all your help!
Ricardo Arguello
----- Original Message -----=20
From: "Thomas Mahler" <tho...@ho...>
To: "Ricardo Arg=FCello" <ri...@ya...>
Cc: "ojb" <obj...@li...>
Sent: Friday, February 01, 2002 2:48 PM
Subject: Re: [OJB-developers] Re: OJB in C/S mode: =
TransactionInProgressEx
Hi Ricardo
attached you'll find the struts sample and a patch for the BrokerPool
(for c/s) mode.
Ricardo Arg=FCello wrote:
> Hello,
>
> As I said before, when 2 users try to use the same part of the
"manager", one of them gets an Exception, the other process it =
correctly.
>
> Maybe it's because of the fact that I'm using ONE "odmg instance"
(odmg=3DOJB.getInstance()) in my manager singleton. In every invocation =
of
a function in this manager, I use:
> Transaction tx =3D odmg.newTransaction()
>
> So, aparently it's not a problem. But I'm not sure...
It's no problem. OJB.getInstance() always returns the same instance...
>
> This is the complete stack trace, when 2 users access the same =
function:
>=20
-------------------------------------------------------------------------=
> [ODMG] ERROR: null
> org.odmg.QueryException
> at ojb.odmg.oql.OQLQueryImpl.execute(OQLQueryImpl.java:251)
> at
charlie.ojb.ProyectoManagerOJB.tareasPorEmpleado(ProyectoManagerOJB.j
> ava)
unfortunately we don't see what's going wrong within the OQL query.
Please set the loglevel for ODMG and PersistenceBrokerImpl to DEBUG and
send me the result to get a clearer impression...
>
>
> I'm NOT using OJB in C/S mode. The "TransactionInProgressException
problem" was related to OJB in C/S mode, this one is ANOTHER problem, as
you can see in the satck trace.
>
> I hope the patch helps solucionate the C/S problem, but this one
seems to be another issue, am I right?
>
Yes, I think so. Please try out the patched c/s mode
HTH,
Thomas
> Thanks for your help!
>
>
> Greetings,
>
>
>
> Ricardo Arguello
> ri...@ya...
>
>
>
>
>
>
> ----- Original Message -----
> From: "Mahler Thomas" <tho...@it...>
> To: "'Ricardo Arg=FCello'" <ri...@ya...>; "ojb"
<obj...@li...>
> Sent: Thursday, January 31, 2002 8:04 AM
> Subject: AW: [OJB-developers] Re: OJB in C/S mode:
TransactionInProgressEx
>
>
> Hi again Ricardo,
>
> The bugfix is on the way. I hope to get it posted today!
>
>
>>-----Urspr=FCngliche Nachricht-----
>>Von: Ricardo Arg=FCello [mailto:ri...@ya...]
>>Gesendet: Mittwoch, 30. Januar 2002 22:43
>>An: ojb
>>Betreff: Re: [OJB-developers] Re: OJB in C/S mode:
>>TransactionInProgressEx
>>
>>
>>Hello,
>>
>>Thanks for your reply.
>>
>>I have a "manager class" wich is called from my Struts
>>actions, and it does al the work, calling the "OJBed
>>classes", retuning Collections, etc.
>>
>>This manager is a singleton, and when I instanciate it (only
>>once) I initialize the "odmg" class variable:
>>
>>public class ProyectoManagerOJB {
>> private static final String databaseName =3D
>>
>>PersistenceBrokerFactory.getConfiguration().getRepositoryFilename();
>>
>> private Implementation odmg; // maybe this is wrong!
>>
>
> This absolutely ok. OJB.getInstance() is a singleton
>
>
>> public ProyectoManagerOJB() {
>> odmg =3D OJB.getInstance();
>> Database db =3D odmg.newDatabase();
>> try {
>> db.open(databaseName, Database.OPEN_READ_WRITE);
>> } catch (ODMGException ex) {
>> ex.printStackTrace();
>> }
>> }
>>...
>>}
>>
>>Struts actions call functions in this class:
>>
>>ProyectoManager pm =3D ProyectoManager.getInstance(); // it's
>>an abstract factory, returns ProyectoManagerOJB
>>User user =3D pm.getUser("john");
>>
>>The corresponding class in ProyectoManagerOJB is this one:
>>(User is an interface, UserOJB implements it)
>>
>>public User getUser(String user) {
>> try {
>> Transaction tx =3D odmg.newTransaction();
>> tx.begin();
>> OQLQuery query =3D odmg.newOQLQuery();
>> query.create(
>> "select p from "
>> + charlie.ojb.UserOJB.class.getName()
>> + " where user =3D \""
>> + user
>> + "\"");
>> DList result =3D (DList) query.execute();
>> User user =3D (User) result.get(0);
>> tx.commit();
>> return user;
>> } catch (Exception e) {
>> e.printStackTrace();
>> return null;
>> }
>>}
>>
>>As you can see, in getUser() I use the class variable "odmg".
>>
>
> That's OK!
>
>
>>If I have two users login in, I get a =
TransactionInProgressException.
>>
>
> Please send me a complete stacktrace of this exception. I'm not sure
if it
> is a PersistenceBrokerException or an ODMGException.
>
>
>>I believe it is because of the fact that odmg is an instance =
variable.
>>
>>So, could you send me some code examples of your
>>friends'"odmg helper class" to see how I could implement it
>>in my architecture?
>>
>>
>
> I'll will post it to you later this day.
>
>
>>Thanks in advance!
>>
>>Greetings from Quito, Ecuador
>>
>>
>
> greetings from Essen, Germany
>
> Thomas
>
>
>>Ricardo Arguello
>>ri...@ya...
>>
>>
>>----- Original Message -----
>>From: "Thomas Mahler" <tho...@ho...>
>>To: "Ricardo Arguello" <ri...@ya...>; "ojb"
>><obj...@li...>
>>Sent: Wednesday, January 30, 2002 3:35 PM
>>Subject: [OJB-developers] Re: OJB in C/S mode: =
TransactionInProgressEx
>>
>>
>>
>>>Hi Ricardo,
>>>
>>>sorry for the trouble OJB is causing for you.
>>> We'll try to get a solution for this soon.
>>>
>>>Ricardo Arguello wrote:
>>>
>>>
>>>>Hello,
>>>>
>>>>As I said before, I'm getting a
>>>>TransactionInProgressException when I use OJB in
>>>>Client/Server mode.
>>>>
>>>>I tried changing the log to DEBUG, bit it imposibe for me
>>>>to distinguish any of the errors, there is a LOT of DEBUG
>>>>messages, of the following style:
>>>>
>>>>
>>>
>>>
>>>Yup, sometimes it's hard to know which part of the system
>>>
>>might write
>>
>>>the relevant log infos...
>>>
>>>
>>>
>>>>[ojb.broker.server.RequestProcessor] DEBUG: OJB
>>>>RequestProcessor 2 waiting...
>>>>[ojb.broker.server.RequestProcessor] DEBUG: OJB
>>>>RequestProcessor 7 working...
>>>>[ojb.broker.server.RequestProcessor] DEBUG: dispatching [24,
>>>>[ojb.broker.server.RequestProcessor] DEBUG: -1,
>>>>[ojb.broker.server.RequestProcessor] DEBUG: class
>>>>charlie.ojb.EmpleadoOJB,
>>>>[ojb.broker.server.RequestProcessor] DEBUG: null]
>>>>[ojb.broker.server.BrokerPool] DEBUG: getFreeBrokerId: 1,
>>>>remaining: [1, 1, 1, 1
>>>>, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 3,
>>>>1, 1, 1, 1, 1, 1, 3,
>>>>1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1, 3, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1,
>>>> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1, 1
>>>>, 1, 1, 4, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 5, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 6, 1,
>>>>1, 1, 1, 1, 1, 1, 1,
>>>> 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 7, 1, 1, 1, 8,
>>>>1, 1, 1, 9, 1, 1, 9, 1
>>>>, 1, 1, 1, 1, 1, 0, 1, 9, 1, 1, 1, 1, 9, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 1, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1,
>>>> 1, 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1, 1
>>>>, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1,
>>>>1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
>>>>
>>>>
>>>>Awful, right?
>>>>
>>>
>>>Yes that's awful to read. But it is also a clear indicator
>>>
>>that there
>>
>>>must be a bug in the BrokerPool. It should contain exactly
>>>
>>ten brokers
>>
>>>numbered from 0 - 9.
>>>As the list you see contains a lot of "1" entries you are
>>>
>>served with
>>
>>>the same broker instance for multiple transactions.
>>>
>>>This is definitely a severe bug.
>>>
>>>I will start working on it asap and hope to have a fix by
>>>
>>the weekend.
>>
>>>
>>>>I cannot find the errors I metioned in my previous
>>>>message... But my application stil acts the same: Two
>>>>persons can't even login at the same time, one of them gets
>>>>a stack trace...
>>>>
>>>>
>>>
>>>I assume that it is all related to the above error !
>>>
>>>
>>>
>>>>I was reading the Locking document, and I tried to add a
>>>>isolation=3D"read-uncommitted" to all my class definitions in
>>>>the repository.xml file, but tha doesn't help either.
>>>>
>>>
>>>This is definitely a locking problem. OJB uses read-uncommited by
>>>default already. In the case of locking problems you would
>>>
>>get a ODMG
>>
>>>LockNotGrantedException!
>>>
>>>
>>>>How
>>>>should I declare my lockings? In the XML file, in my code
>>>>(you say the ODMG doesnt dictate a way to that).
>>>>
>>>>
>>>
>>>Locking is declared in the repository only. ODMG specifies
>>>
>>no API for
>>
>>>setting locklevels.
>>>
>>>
>>>
>>>>Please clarify this to me? Do I need to do C/S to use OJB
>>>>in a web application?
>>>>
>>>
>>>
>>>Hmm. I know of people who are using OJB in singlevm mode in
>>>
>>Web apps.
>>
>>>As long as the commit phase of ODMG transactions is
>>>
>>relatively short
>>
>>>there won't be much problems.
>>>
>>>I'm intending to have brokerpool for the singlevm mode too.
>>>
>>then it will
>>
>>> be totally safe to use the singlevm mode which is much faster !!!
>>>
>>>
>>>
>>>>Wouldn't a "OJB Helper" with a single Database instance
>>>>(singleton) be enough?
>>>>
>>>>
>>>
>>>Yes that should be sufficient.
>>>
>>>
>>>
>>>>If that's the case, how could I implement that?
>>>>
>>>
>>>Hmm, I'm not quite sure what you want to implement exactly?
>>>
>>>
>>>
>>>>Is OJB thread safe?
>>>>
>>>
>>>Yes
>>>
>>>
>>>>How can I use OJB with Tomcat + Struts?
>>>>
>>>
>>>A colleague of mine used Struts and the OJB ODMG api sucessfully.
>>>He defined special STruts actions that used ODMG
>>>
>>transactions to do the
>>
>>>persistence work. It's pretty straightforward.
>>>
>>>If you are interested I might post you sample code for this.
>>>
>>>
>>>
>>>>How can I assure that 2 transactions can READ at the same
>>>>time, without locking?
>>>>
>>>>
>>>
>>>Currently the OJB ODMG implementation relies on pessimistic =
locking.
>>>There is currently no optimistic locking support. (will be
>>>
>>there in 1.0)
>>
>>>But if you use "uncommitted-read" there should be no
>>>
>>locking problems.
>>
>>>You can do even more do avoid lock conflicts through WRITE
>>>
>>operations.
>>
>>>In OJB.properties you'll find an entry LockAssociations set
>>>
>>this to READ.
>>
>>>The LockAssociations entry defines the behaviour for the OJB
>>>implicit locking feature. If set to WRITE (default)
>>>
>>acquiring a write-
>>
>>>lock on a given object x implies write locks on all objects
>>>
>>associated
>>
>>>to x. If set to READ implicit read-locks are acquired.
>>>Acquiring a read-lock on x thus allways results in implicit
>>>
>>read-locks
>>
>>>on all associated objects.
>>>
>>>Using uncommited reads in combination with
>>>
>>LockAssociation=3DREAD will
>>
>>>reduce Locking related problems to a minimum.
>>>
>>>But as I said before: your problem is NOT related to locking.
>>>
>>>
>>>
>>>>Thanks in advance.
>>>>
>>>>Ricardo Arguello (an ex-Castor refugee)
>>>>
>>>
>>>I'll work hard to ensure you won't become an OJB refugee ;-)
>>>
>>>
>>>cu,
>>>
>>>Thomas
>>>
>>>
>>>
>>>_______________________________________________
>>>Objectbridge-developers mailing list
>>>Obj...@li...
=
>>>https://lists.sourceforge.net/lists/listinfo/objectbridge-developers
>>>
>>>
>>
>>
>>_______________________________________________
>>Objectbridge-developers mailing list
>>Obj...@li...
>>https://lists.sourceforge.net/lists/listinfo/objectbridge-developers
>>
>>
>
> _______________________________________________
> Objectbridge-developers mailing list
> Obj...@li...
> https://lists.sourceforge.net/lists/listinfo/objectbridge-developers
>
>
> _______________________________________________
> Objectbridge-developers mailing list
> Obj...@li...
> https://lists.sourceforge.net/lists/listinfo/objectbridge-developers
>
>
>
>
|
