You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@activemq.apache.org by Michael André Pearce <mi...@me.com> on 2018/03/07 05:00:39 UTC
Fwd: ActiveMq Artemis crashes when large amqp messages need are
delivered
@Clebert
Could this be a blocker for release?
Just sending it to the dev mail list incase it got missed in the users.
Begin forwarded message:
> From: andi welchlin <an...@gmail.com>
> Date: 6 March 2018 at 09:08:39 GMT
> To: users@activemq.apache.org
> Subject: ActiveMq Artemis crashes when large amqp messages need are delivered
> Reply-To: users@activemq.apache.org
>
> Hello,
>
> I tested Artemis as a stanalone broker (the current snapshot 2.5.0).
>
> One program sent a 100MB amqp message to a queue in the broker. This could be done.
>
> But while the message was delivered to a subscriber Artemis crashed.
>
> The queue was defined like this:
>
> <queues>
> <queue name="awe.test.queue">
> <address>awe.test.queue</address>
> <durable>true</durable>
> </queue>
> </queues>
>
> And I set global size to 1000 MB:
>
> <global-max-size>1000Mb</global-max-size>
>
>
>
> When I restarted the broker and connected the subscriber it crashed, again.
>
> Since the log output is large I attached a file to this mail. The end of the output looks like this:
>
> at java.lang.Object.wait(Native Method)
> - waiting on java.lang.ref.ReferenceQueue$Lock@773bc0c1
> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:143)
> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:164)
> at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)
>
>
> "Reference Handler" Id=2 WAITING on java.lang.ref.Reference$Lock@3e627566
> at java.lang.Object.wait(Native Method)
> - waiting on java.lang.ref.Reference$Lock@3e627566
> at java.lang.Object.wait(Object.java:502)
> at java.lang.ref.Reference.tryHandlePending(Reference.java:191)
> at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)
>
>
> ===============================================================================
> End Thread dump
> *******************************************************************************
>
>
> I tested the same setup with a QPID C++ broker. It had no problems even with much larger messages.
>
> Kind Regards,
> Andreas
>
>
>
>
Re: ActiveMq Artemis crashes when large amqp messages need are delivered
Posted by Clebert Suconic <cl...@gmail.com>.
My bad.. wrong list :)
On Wed, Mar 7, 2018 at 3:01 PM, Michael André Pearce
<mi...@me.com> wrote:
> I don’t , i was forwarding on something raised in the user mail list, just sounded important. Probably worth replying back to them in the user mail list
>
> Sent from my iPhone
>
>> On 7 Mar 2018, at 19:36, Clebert Suconic <cl...@gmail.com> wrote:
>>
>> If you need really large messages.. can you use the core protocol clients?
>>
>> On Wed, Mar 7, 2018 at 8:30 AM, Clebert Suconic
>> <cl...@gmail.com> wrote:
>>> I don’t think so. We don’t stream large messages in AMQP yet. We just
>>> convert them as a single chunk. It should work as long as you have memory.
>>>
>>> Netty uses native buffers. So there is a chance this could crash when OME.
>>>
>>>
>>> The other issue I saw on users list seems a blocker and I will be looking on
>>> that.
>>>
>>> On Wed, Mar 7, 2018 at 12:01 AM Michael André Pearce
>>> <mi...@me.com> wrote:
>>>>
>>>> @Clebert
>>>>
>>>> Could this be a blocker for release?
>>>> Just sending it to the dev mail list incase it got missed in the users.
>>>>
>>>> Begin forwarded message:
>>>>
>>>> From: andi welchlin <an...@gmail.com>
>>>> Date: 6 March 2018 at 09:08:39 GMT
>>>> To: users@activemq.apache.org
>>>> Subject: ActiveMq Artemis crashes when large amqp messages need are
>>>> delivered
>>>> Reply-To: users@activemq.apache.org
>>>>
>>>> Hello,
>>>>
>>>> I tested Artemis as a stanalone broker (the current snapshot 2.5.0).
>>>>
>>>> One program sent a 100MB amqp message to a queue in the broker. This could
>>>> be done.
>>>>
>>>> But while the message was delivered to a subscriber Artemis crashed.
>>>>
>>>> The queue was defined like this:
>>>>
>>>> <queues>
>>>> <queue name="awe.test.queue">
>>>> <address>awe.test.queue</address>
>>>> <durable>true</durable>
>>>> </queue>
>>>> </queues>
>>>>
>>>> And I set global size to 1000 MB:
>>>>
>>>> <global-max-size>1000Mb</global-max-size>
>>>>
>>>>
>>>>
>>>> When I restarted the broker and connected the subscriber it crashed,
>>>> again.
>>>>
>>>> Since the log output is large I attached a file to this mail. The end of
>>>> the output looks like this:
>>>>
>>>> at java.lang.Object.wait(Native Method)
>>>> - waiting on java.lang.ref.ReferenceQueue$Lock@773bc0c1
>>>> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:143)
>>>> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:164)
>>>> at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)
>>>>
>>>>
>>>> "Reference Handler" Id=2 WAITING on java.lang.ref.Reference$Lock@3e627566
>>>> at java.lang.Object.wait(Native Method)
>>>> - waiting on java.lang.ref.Reference$Lock@3e627566
>>>> at java.lang.Object.wait(Object.java:502)
>>>> at java.lang.ref.Reference.tryHandlePending(Reference.java:191)
>>>> at
>>>> java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)
>>>>
>>>>
>>>>
>>>> ===============================================================================
>>>> End Thread dump
>>>>
>>>> *******************************************************************************
>>>>
>>>>
>>>> I tested the same setup with a QPID C++ broker. It had no problems even
>>>> with much larger messages.
>>>>
>>>> Kind Regards,
>>>> Andreas
>>>>
>>>>
>>>>
>>>>
>>> --
>>> Clebert Suconic
>>
>>
>>
>> --
>> Clebert Suconic
--
Clebert Suconic
Re: ActiveMq Artemis crashes when large amqp messages need are
delivered
Posted by Michael André Pearce <mi...@me.com>.
I don’t , i was forwarding on something raised in the user mail list, just sounded important. Probably worth replying back to them in the user mail list
Sent from my iPhone
> On 7 Mar 2018, at 19:36, Clebert Suconic <cl...@gmail.com> wrote:
>
> If you need really large messages.. can you use the core protocol clients?
>
> On Wed, Mar 7, 2018 at 8:30 AM, Clebert Suconic
> <cl...@gmail.com> wrote:
>> I don’t think so. We don’t stream large messages in AMQP yet. We just
>> convert them as a single chunk. It should work as long as you have memory.
>>
>> Netty uses native buffers. So there is a chance this could crash when OME.
>>
>>
>> The other issue I saw on users list seems a blocker and I will be looking on
>> that.
>>
>> On Wed, Mar 7, 2018 at 12:01 AM Michael André Pearce
>> <mi...@me.com> wrote:
>>>
>>> @Clebert
>>>
>>> Could this be a blocker for release?
>>> Just sending it to the dev mail list incase it got missed in the users.
>>>
>>> Begin forwarded message:
>>>
>>> From: andi welchlin <an...@gmail.com>
>>> Date: 6 March 2018 at 09:08:39 GMT
>>> To: users@activemq.apache.org
>>> Subject: ActiveMq Artemis crashes when large amqp messages need are
>>> delivered
>>> Reply-To: users@activemq.apache.org
>>>
>>> Hello,
>>>
>>> I tested Artemis as a stanalone broker (the current snapshot 2.5.0).
>>>
>>> One program sent a 100MB amqp message to a queue in the broker. This could
>>> be done.
>>>
>>> But while the message was delivered to a subscriber Artemis crashed.
>>>
>>> The queue was defined like this:
>>>
>>> <queues>
>>> <queue name="awe.test.queue">
>>> <address>awe.test.queue</address>
>>> <durable>true</durable>
>>> </queue>
>>> </queues>
>>>
>>> And I set global size to 1000 MB:
>>>
>>> <global-max-size>1000Mb</global-max-size>
>>>
>>>
>>>
>>> When I restarted the broker and connected the subscriber it crashed,
>>> again.
>>>
>>> Since the log output is large I attached a file to this mail. The end of
>>> the output looks like this:
>>>
>>> at java.lang.Object.wait(Native Method)
>>> - waiting on java.lang.ref.ReferenceQueue$Lock@773bc0c1
>>> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:143)
>>> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:164)
>>> at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)
>>>
>>>
>>> "Reference Handler" Id=2 WAITING on java.lang.ref.Reference$Lock@3e627566
>>> at java.lang.Object.wait(Native Method)
>>> - waiting on java.lang.ref.Reference$Lock@3e627566
>>> at java.lang.Object.wait(Object.java:502)
>>> at java.lang.ref.Reference.tryHandlePending(Reference.java:191)
>>> at
>>> java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)
>>>
>>>
>>>
>>> ===============================================================================
>>> End Thread dump
>>>
>>> *******************************************************************************
>>>
>>>
>>> I tested the same setup with a QPID C++ broker. It had no problems even
>>> with much larger messages.
>>>
>>> Kind Regards,
>>> Andreas
>>>
>>>
>>>
>>>
>> --
>> Clebert Suconic
>
>
>
> --
> Clebert Suconic
Re: ActiveMq Artemis crashes when large amqp messages need are delivered
Posted by Clebert Suconic <cl...@gmail.com>.
If you need really large messages.. can you use the core protocol clients?
On Wed, Mar 7, 2018 at 8:30 AM, Clebert Suconic
<cl...@gmail.com> wrote:
> I don’t think so. We don’t stream large messages in AMQP yet. We just
> convert them as a single chunk. It should work as long as you have memory.
>
> Netty uses native buffers. So there is a chance this could crash when OME.
>
>
> The other issue I saw on users list seems a blocker and I will be looking on
> that.
>
> On Wed, Mar 7, 2018 at 12:01 AM Michael André Pearce
> <mi...@me.com> wrote:
>>
>> @Clebert
>>
>> Could this be a blocker for release?
>> Just sending it to the dev mail list incase it got missed in the users.
>>
>> Begin forwarded message:
>>
>> From: andi welchlin <an...@gmail.com>
>> Date: 6 March 2018 at 09:08:39 GMT
>> To: users@activemq.apache.org
>> Subject: ActiveMq Artemis crashes when large amqp messages need are
>> delivered
>> Reply-To: users@activemq.apache.org
>>
>> Hello,
>>
>> I tested Artemis as a stanalone broker (the current snapshot 2.5.0).
>>
>> One program sent a 100MB amqp message to a queue in the broker. This could
>> be done.
>>
>> But while the message was delivered to a subscriber Artemis crashed.
>>
>> The queue was defined like this:
>>
>> <queues>
>> <queue name="awe.test.queue">
>> <address>awe.test.queue</address>
>> <durable>true</durable>
>> </queue>
>> </queues>
>>
>> And I set global size to 1000 MB:
>>
>> <global-max-size>1000Mb</global-max-size>
>>
>>
>>
>> When I restarted the broker and connected the subscriber it crashed,
>> again.
>>
>> Since the log output is large I attached a file to this mail. The end of
>> the output looks like this:
>>
>> at java.lang.Object.wait(Native Method)
>> - waiting on java.lang.ref.ReferenceQueue$Lock@773bc0c1
>> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:143)
>> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:164)
>> at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)
>>
>>
>> "Reference Handler" Id=2 WAITING on java.lang.ref.Reference$Lock@3e627566
>> at java.lang.Object.wait(Native Method)
>> - waiting on java.lang.ref.Reference$Lock@3e627566
>> at java.lang.Object.wait(Object.java:502)
>> at java.lang.ref.Reference.tryHandlePending(Reference.java:191)
>> at
>> java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)
>>
>>
>>
>> ===============================================================================
>> End Thread dump
>>
>> *******************************************************************************
>>
>>
>> I tested the same setup with a QPID C++ broker. It had no problems even
>> with much larger messages.
>>
>> Kind Regards,
>> Andreas
>>
>>
>>
>>
> --
> Clebert Suconic
--
Clebert Suconic
Re: ActiveMq Artemis crashes when large amqp messages need are delivered
Posted by Clebert Suconic <cl...@gmail.com>.
I don’t think so. We don’t stream large messages in AMQP yet. We just
convert them as a single chunk. It should work as long as you have memory.
Netty uses native buffers. So there is a chance this could crash when OME.
The other issue I saw on users list seems a blocker and I will be looking
on that.
On Wed, Mar 7, 2018 at 12:01 AM Michael André Pearce <
michael.andre.pearce@me.com> wrote:
> @Clebert
>
> Could this be a blocker for release?
> Just sending it to the dev mail list incase it got missed in the users.
>
> Begin forwarded message:
>
> *From:* andi welchlin <an...@gmail.com>
> *Date:* 6 March 2018 at 09:08:39 GMT
> *To:* users@activemq.apache.org
> *Subject:* *ActiveMq Artemis crashes when large amqp messages need are
> delivered*
> *Reply-To:* users@activemq.apache.org
>
> Hello,
>
> I tested Artemis as a stanalone broker (the current snapshot 2.5.0).
>
> One program sent a 100MB amqp message to a queue in the broker. This could
> be done.
>
> But while the message was delivered to a subscriber Artemis crashed.
>
> The queue was defined like this:
>
> <queues>
> <queue name="awe.test.queue">
> <address>awe.test.queue</address>
> <durable>true</durable>
> </queue>
> </queues>
>
> And I set global size to 1000 MB:
>
> <global-max-size>1000Mb</global-max-size>
>
>
>
> When I restarted the broker and connected the subscriber it crashed, again.
>
> Since the log output is large I attached a file to this mail. The end of
> the output looks like this:
>
> at java.lang.Object.wait(Native Method)
> - waiting on java.lang.ref.ReferenceQueue$Lock@773bc0c1
> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:143)
> at java.lang.ref.ReferenceQueue.remove(ReferenceQueue.java:164)
> at java.lang.ref.Finalizer$FinalizerThread.run(Finalizer.java:209)
>
>
> "Reference Handler" Id=2 WAITING on java.lang.ref.Reference$Lock@3e627566
> at java.lang.Object.wait(Native Method)
> - waiting on java.lang.ref.Reference$Lock@3e627566
> at java.lang.Object.wait(Object.java:502)
> at java.lang.ref.Reference.tryHandlePending(Reference.java:191)
> at java.lang.ref.Reference$ReferenceHandler.run(Reference.java:153)
>
>
>
> ===============================================================================
> End Thread dump
>
> *******************************************************************************
>
>
> I tested the same setup with a QPID C++ broker. It had no problems even
> with much larger messages.
>
> Kind Regards,
> Andreas
>
>
>
>
> --
Clebert Suconic