Hello Gerald,
I completely agree. I guess, one could consider this a bug (but of
course, by definition, BaseX does not has bugs; it's hidden features).
Seems like something is going wrong when releasing the write lock. Could
you please test it with our latest snapshot, to see if the issue was
already fixed?
Christian is currently on vacation, so an actual bug fix could take some
more time than usual. If I have time, I'll try to look into it.
Cheers,
Dirk
Dirk Kirsten, BaseX GmbH, http://basex.org
On 09/24/2014 05:41 PM, Gerald de Jong wrote:
> It happened again. Could I not argue that nothing I can do as a client to
> BaseX should be able to freeze it? I mean, even if, for some reason,
> things are getting muddled up.
>
> Exception in thread "Thread-56391" java.lang.IllegalMonitorStateException
> at
> java.util.concurrent.locks.ReentrantReadWriteLock$Sync.tryRelease(ReentrantReadWriteLock.java:374)
> at
> java.util.concurrent.locks.AbstractQueuedSynchronizer.release(AbstractQueuedSynchronizer.java:1260)
> at
> java.util.concurrent.locks.ReentrantReadWriteLock$WriteLock.unlock(ReentrantReadWriteLock.java:1131)
> at org.basex.core.DBLocking.release(DBLocking.java:219)
> at org.basex.core.Context.unregister(Context.java:249)
> at org.basex.core.Command.execute(Command.java:104)
>
>
> On Tue, Sep 23, 2014 at 2:17 PM, Christian Grün <christian.gruen@gmail.com>
> wrote:
>
>> You can set the debugging flag and check out the database log files to
>> get more insight into what's going on.
>>
>>
>> On Tue, Sep 23, 2014 at 2:15 PM, Gerald de Jong <gerald@delving.eu> wrote:
>>> Can I start up the basexserver with some kind of logging so I can see
>> what
>>> leads to the IllegalMonitorStateException?
>>>
>>> On Tue, Sep 23, 2014 at 1:53 PM, Christian Grün <
>> christian.gruen@gmail.com>
>>> wrote:
>>>>
>>>>> Looking at these tests, it appears that none of them do large numbers
>> of
>>>>> ADD
>>>>> commands. Did I miss something?
>>>>
>>>> You may need to check the other tests. But as the (updating) add
>>>> command will never be executed in parallel, it shouldn't cause any
>>>> problems with parallelization (feel free to tell me if I'm wrong).
>>>>
>>>>
>>>>>
>>>>> On Tue, Sep 23, 2014 at 1:34 PM, Christian Grün
>>>>> <christian.gruen@gmail.com>
>>>>> wrote:
>>>>>>
>>>>>>> The fact that it's scala should be insignificant, since I'm using
>> the
>>>>>>> Java
>>>>>>> API.
>>>>>>
>>>>>> Well, it might be theoretically insignificant, but we cannot give a
>>>>>> quick answer on what's your problem. In general, parallel client
>>>>>> requests are no problem, we have various JUnit tests doing exactly
>>>>>> that (see e.g. [1], [2] or [3]).
>>>>>>
>>>>>> [1]
>>>>>>
>>>>>>
>> https://github.com/BaseXdb/basex/blob/master/basex-core/src/test/java/org/basex/server/ConcurrencyTest.java
>>>>>> [2]
>>>>>>
>>>>>>
>> https://github.com/BaseXdb/basex/blob/master/basex-tests/src/test/java/org/basex/performance/RESTConcurrencyTest.java
>>>>>> [3]
>>>>>>
>>>>>>
>> https://github.com/BaseXdb/basex/blob/master/basex-tests/src/test/java/org/basex/performance/ServerStressTest.java
>>>>>>
>>>>>>
>>>>>>>
>>>>>>> What's happening is that two (or more) threads are doing this at
>> the
>>>>>>> same
>>>>>>> time:
>>>>>>>
>>>>>>> ClientSession session = new ClientSession(...);
>>>>>>> session.execute(new Open(database));
>>>>>>> session.execute(new Set("autoflush", "false"));
>>>>>>> session.add([path string], [document-bytes]); // a zillion
>> times
>>>>>>> session.execute(new Flush());
>>>>>>> session.close();
>>>>>>>
>>>>>>> My first question is of course if this should work parallel with no
>>>>>>> problems
>>>>>>> "in theory".
>>>>>>>
>>>>>>> On Tue, Sep 23, 2014 at 12:46 PM, Christian Grün
>>>>>>> <christian.gruen@gmail.com>
>>>>>>> wrote:
>>>>>>>>
>>>>>>>> ...difficult to tell, as we don't have too much experience with
>>>>>>>> Scala.
>>>>>>>> Could you provide this example as Java code?
>>>>>>>>
>>>>>>>>
>>>>>>>>
>>>>>>>> On Tue, Sep 23, 2014 at 12:10 PM, Gerald de Jong <
>> gerald@delving.eu>
>>>>>>>> wrote:
>>>>>>>>> Dear BaseX experts,
>>>>>>>>>
>>>>>>>>> I'm wondering if anybody can see something that I'm doing
>> wrong. I
>>>>>>>>> would
>>>>>>>>> expect it to be possible to add documents from multiple threads.
>>>>>>>>> Am
>>>>>>>>> I
>>>>>>>>> wrong
>>>>>>>>> about that??
>>>>>>>>>
>>>>>>>>> When I set two of my datasets saving documents to BaseX 7.9 at
>> the
>>>>>>>>> same
>>>>>>>>> time, I get this exception:
>>>>>>>>>
>>>>>>>>> $ basexserver
>>>>>>>>> BaseX 7.9 [Server]
>>>>>>>>> Server was started (port: 1984)
>>>>>>>>> Exception in thread "Thread-2295"
>>>>>>>>> java.lang.IllegalMonitorStateException
>>>>>>>>> at
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>> java.util.concurrent.locks.ReentrantReadWriteLock$Sync.tryRelease(ReentrantReadWriteLock.java:374)
>>>>>>>>> at
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>> java.util.concurrent.locks.AbstractQueuedSynchronizer.release(AbstractQueuedSynchronizer.java:1260)
>>>>>>>>> at
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>>>>>>>>>
>> java.util.concurrent.locks.ReentrantReadWriteLock$WriteLock.unlock(ReentrantReadWriteLock.java:1131)
>>>>>>>>> at org.basex.core.DBLocking.release(DBLocking.java:219)
>>>>>>>>> at org.basex.core.Context.unregister(Context.java:249)
>>>>>>>>> at org.basex.core.Command.execute(Command.java:104)
>>>>>>>>> at org.basex.core.Command.execute(Command.java:117)
>>>>>>>>> at
>>>>>>>>> org.basex.server.ClientListener.execute(ClientListener.java:384)
>>>>>>>>> at org.basex.server.ClientListener.add(ClientListener.java:355)
>>>>>>>>> at org.basex.server.ClientListener.run(ClientListener.java:100)
>>>>>>>>>
>>>>>>>>> All I am doing is making concurrent calls in two parallel
>>>>>>>>> sessions:
>>>>>>>>>
>>>>>>>>> session = new ClientSession(...)
>>>>>>>>> session.execute(new Open(database))
>>>>>>>>> session.execute(new Set("autoflush", "false"))
>>>>>>>>> session.add([path string], [document-bytes]) // a zillion
>>>>>>>>> times
>>>>>>>>> session.execute(new Flush())
>>>>>>>>>
>>>>>>>>> During the execution of the parallel add commands, it crashes.
>>>>>>>>>
>>>>>>>>> The relevant code:
>>>>>>>>>
>>>>>>>>> def withSession[T](block: ClientSession => T): T = {
>>>>>>>>> val session = new ClientSession(host, port, user, pass)
>>>>>>>>> try {
>>>>>>>>> block(session)
>>>>>>>>> }
>>>>>>>>> finally {
>>>>>>>>> session.close()
>>>>>>>>> }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> def withDbSession[T](database: String)(block: ClientSession =>
>>>>>>>>> T):
>>>>>>>>> T =
>>>>>>>>> {
>>>>>>>>> withSession {
>>>>>>>>> session =>
>>>>>>>>> session.execute(new Open(database))
>>>>>>>>> session.execute(new Set("autoflush", "false"))
>>>>>>>>> val result = block(session)
>>>>>>>>> session.execute(new Flush())
>>>>>>>>> result
>>>>>>>>> }
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> then:
>>>>>>>>>
>>>>>>>>> withDbSession {
>>>>>>>>> session =>
>>>>>>>>> // many calls to session.add(...)
>>>>>>>>> }
>>>>>>>>>
>>>>>>>>> --
>>>>>>>>> Delving BV, Vasteland 8, Rotterdam
>>>>>>>>> http://www.delving.eu
>>>>>>>>> http://twitter.com/fluxe
>>>>>>>>> skype: beautifulcode
>>>>>>>>> +31629339805
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>
>>>>>>> --
>>>>>>> Delving BV, Vasteland 8, Rotterdam
>>>>>>> http://www.delving.eu
>>>>>>> http://twitter.com/fluxe
>>>>>>> skype: beautifulcode
>>>>>>> +31629339805
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> --
>>>>> Delving BV, Vasteland 8, Rotterdam
>>>>> http://www.delving.eu
>>>>> http://twitter.com/fluxe
>>>>> skype: beautifulcode
>>>>> +31629339805
>>>
>>>
>>>
>>>
>>> --
>>> Delving BV, Vasteland 8, Rotterdam
>>> http://www.delving.eu
>>> http://twitter.com/fluxe
>>> skype: beautifulcode
>>> +31629339805
>>
>
>
>
--
|-- Firmensitz: Blarerstrasse 56, 78462 Konstanz
|-- Registergericht Freiburg, HRB: 708285, Geschäftsführer:
| Dr. Christian Grün, Dr. Alexander Holupirek, Michael Seiferle
`-- Phone: 0049 7531 28 28 676, Fax: 0049 7531 20 05 22