You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@activemq.apache.org by Rainer Klute <kl...@rainer-klute.de> on 2007/01/24 19:37:46 UTC
OutOfMemoryErrors again
Rob Davies wrote in another thread:
> it would be great if you could produce a junit test case and attach it
> to an issue in our issue tracker -
> http://issues.apache.org/activemq/secure/CreateIssue!default.jspa
> these are just the sort of cases we look for for testing n' fixing stuff!
JUnit test cases are great. However, I am not sure whether it is
possible to write them for the various OutOfMemoryErrors I encounter.
Such an error can occur everywhere if memory is tight and you cannot
ensure you can catch it and handle it properly.
Currently I encouter OutOfMemoryErrors (OOME) with the kaha persistence
manager when consuming messages from a topic. My test program writes a
lot of small messages (say 10 million) to a topic and consumes them in
another thread. I timed the two threads in a way that the consuming
thread kicks in only after the producing thread finished its work. The
result: The producer delivers the messages to the topic very fast.
However, while the consumer consumes, the application needs more and
more memory and more and more time for garbage collection to free less
and less space. Finally it gets stuck with an OOME. Could it be that the
persistence layer keeps references to the consumed messages?
My requirements are roughly 260 more or less small messages per second
with the topic being able to buffer messages for days to weeks.
Best regards
Rainer Klute
Rainer Klute IT-Consulting GmbH
Dipl.-Inform.
Rainer Klute E-Mail: klute@rainer-klute.de
Körner Grund 24 Telefon: +49 172 2324824
D-44143 Dortmund Telefax: +49 231 5349423
OpenPGP fingerprint: E4E4386515EE0BED5C162FBB5343461584B5A42E
Re: OutOfMemoryErrors again
Posted by Rob Davies <ra...@gmail.com>.
Hi Albert,
yes they will - its all the same underlying issue
cheers,
Rob
On 24 Jan 2007, at 18:53, Albert Strasheim wrote:
> Hello all
>
> On Wed, 24 Jan 2007, Rob Davies wrote:
>
>> I'm fixing these at the moment - though I'd expect it will be closer
>> to the end of the week when I've finished testing before committing
>> something
>
> Rob, will these fixes also address the problems I reported earlier
> this
> week with producers getting blocked on UsageManager.waitForSpace? As I
> think I mentioned in that thread, the few times the producer didn't
> block, one would see an OutOfMemoryError fly past every now and then.
>
> Thanks!
>
> Cheers,
>
> Albert
Re: OutOfMemoryErrors again
Posted by Albert Strasheim <fu...@gmail.com>.
Hello all
On Wed, 24 Jan 2007, Rob Davies wrote:
> I'm fixing these at the moment - though I'd expect it will be closer
> to the end of the week when I've finished testing before committing
> something
Rob, will these fixes also address the problems I reported earlier this
week with producers getting blocked on UsageManager.waitForSpace? As I
think I mentioned in that thread, the few times the producer didn't
block, one would see an OutOfMemoryError fly past every now and then.
Thanks!
Cheers,
Albert
Re: OutOfMemoryErrors again
Posted by goldi <go...@freenet.de>.
Hi everybody,
has this problem already been fixed? I also getting OOME with kaha
persistence adapter.
Greets Goldi
Rainer Klute (Rainer Klute IT-Consulting) wrote:
>
> Rob Davies schrieb:
>> I'm fixing these at the moment - though I'd expect it will be closer
>> to the end of the week when I've finished testing before committing
>> something
>
> That's very good news! I am looking forward for a fixed snapshot and
> wish you all the best in your efforts!
>
> Best regards
> Rainer Klute
>
> Rainer Klute IT-Consulting GmbH
> Dipl.-Inform.
> Rainer Klute E-Mail: klute@rainer-klute.de
> Körner Grund 24 Telefon: +49 172 2324824
> D-44143 Dortmund Telefax: +49 231 5349423
>
> OpenPGP fingerprint: E4E4386515EE0BED5C162FBB5343461584B5A42E
>
>
>
>
>
--
View this message in context: http://www.nabble.com/OutOfMemoryErrors-again-tf3083798s2354.html#a11414833
Sent from the ActiveMQ - User mailing list archive at Nabble.com.
Re: OutOfMemoryErrors again
Posted by Rainer Klute <kl...@rainer-klute.de>.
Rob Davies schrieb:
> I'm fixing these at the moment - though I'd expect it will be closer
> to the end of the week when I've finished testing before committing
> something
That's very good news! I am looking forward for a fixed snapshot and
wish you all the best in your efforts!
Best regards
Rainer Klute
Rainer Klute IT-Consulting GmbH
Dipl.-Inform.
Rainer Klute E-Mail: klute@rainer-klute.de
Körner Grund 24 Telefon: +49 172 2324824
D-44143 Dortmund Telefax: +49 231 5349423
OpenPGP fingerprint: E4E4386515EE0BED5C162FBB5343461584B5A42E
Re: OutOfMemoryErrors again
Posted by Rob Davies <ra...@gmail.com>.
I'm fixing these at the moment - though I'd expect it will be closer
to the end of the week when I've finished testing before committing
something
cheers,
Rob
On 24 Jan 2007, at 18:37, Rainer Klute wrote:
> Rob Davies wrote in another thread:
>> it would be great if you could produce a junit test case and
>> attach it
>> to an issue in our issue tracker -
>> http://issues.apache.org/activemq/secure/CreateIssue!default.jspa
>> these are just the sort of cases we look for for testing n' fixing
>> stuff!
>
> JUnit test cases are great. However, I am not sure whether it is
> possible to write them for the various OutOfMemoryErrors I encounter.
> Such an error can occur everywhere if memory is tight and you cannot
> ensure you can catch it and handle it properly.
>
> Currently I encouter OutOfMemoryErrors (OOME) with the kaha
> persistence
> manager when consuming messages from a topic. My test program writes a
> lot of small messages (say 10 million) to a topic and consumes them in
> another thread. I timed the two threads in a way that the consuming
> thread kicks in only after the producing thread finished its work. The
> result: The producer delivers the messages to the topic very fast.
> However, while the consumer consumes, the application needs more and
> more memory and more and more time for garbage collection to free less
> and less space. Finally it gets stuck with an OOME. Could it be
> that the
> persistence layer keeps references to the consumed messages?
>
> My requirements are roughly 260 more or less small messages per second
> with the topic being able to buffer messages for days to weeks.
>
> Best regards
> Rainer Klute
>
> Rainer Klute IT-Consulting GmbH
> Dipl.-Inform.
> Rainer Klute E-Mail: klute@rainer-klute.de
> Körner Grund 24 Telefon: +49 172 2324824
> D-44143 Dortmund Telefax: +49 231 5349423
>
> OpenPGP fingerprint: E4E4386515EE0BED5C162FBB5343461584B5A42E
>