gt;
>
>
>
>
> On 8/21/15, 5:28 AM, "tbai...@gmail.com on behalf of Tim Bain" <
> tbai...@gmail.com on behalf of tb...@alumni.duke.edu> wrote:
>
> >You can tell if a consumer is slower than the rest via JMX. Look at each
> >subscription for the topic an
What do you mean by "cluster"? Network of brokers? Master/slave pair?
Something else?
What's your configuration? Your persistence store?
Thanks for including the version number, but you left out a lot of other
information if you want advice about performance (which is the hardest
thing to give
ons come from. Don't
> spend a lot of time trying to figure out what happened in version 5.2, it's
> totally appropriate that I should be on the hook to upgrade to a modern
> version.
>
> And thanks for the confirmation. :)
>
>
>
>
>
> On 8
Next time it happens, see if you can browse the messages on the queues.
That might indicate whether the problem is the data store (if you can't
browse the queues) or something more localized like the dispatch thread (if
you can).
Also, you need to take thread dumps and/or use a profiler (both is b
Your earlier descriptions sounded like PFC kicked in for the whole-store
limit but not the per-destination limit, which would have been a bug. Does
this email mean that it's kicking in for the per-destination limit after
all?
I have wiki edit rights and I've also felt that the PFC documentation
w
I always thought the broker's throughput was limited more by the number of
messages per second than by the number of connections, but I've never tried
connecting 1000+ consumers so I don't have personal experience.
Are you persisting the messages to a persistence store? You'll get lower
throughpu
I believe you said that PFC kicked in based on the per-destination limit
without setting ProducerWindowSize; is that accurate? If so, I'm assuming
it was whole-connection PFC. If you wanted per-producer PFC, I expect you
would need to set ProducerWindowSize, though I've never run with batches
and
Your config snippet is also missing a closing quote on the wss URI; is that
just an artifact of posting your question?
On Aug 27, 2015 6:02 AM, "Christopher Shannon" <
christopher.l.shan...@gmail.com> wrote:
> You should just use wss://myurl:61616 for your URL. I don't believe you
> need the stom
Have you made config changes to allow the broker to trust your self-signed
cert?
On Aug 27, 2015 1:17 PM, "ArturoBelano" wrote:
> I tried to connect using the openssl client and got:
> connect: Connection refused
> connect:errno=61
> That doesn't narrow the issue down too much for me though. I'm
Why have you configured 10.80.1.1 to connect to itself, and why are 2 and 3
default configurations?
More generally, what are you trying to accomplish here?
I assume you've read the content at
http://activemq.apache.org/masterslave.html, but if not, you need to.
On Sep 1, 2015 3:25 AM, "wang" wro
James,
There are two independent concepts here:
1. what limit gets evaluated to determine whether the producer should stop
sending messages
2. which producers should be stopped in response
Without ProducerWindowSize, the limits are based on the destination and the
entire store (plus other limits
network connection from vm://10.80.1.3?async=false
> to failover:(tcp://10.80.1.1:61616,tcp://10.80.1.2:61616
> ,tcp://10.80.1.3:61616)?randomize=false&maxReconnectAttempts=0
> INFO | 10.80.1.3 bridge to 10.80.1.1 stopped
> INFO | Successfully connected to tcp://10.80.1.1:61616
they were sent). But I donot know how to configure a
> network of brokers, so I tried to use master/slave configure.
>
> I want a network of brokers working with load balancer with single
> outside IP address.
>
> Thanks,
> Wang
>
> On 2015年09月01日 23:05, Tim Bain wrote:
&
server mqtt_02 10.80.1.2:1883 check
> server mqtt_03 10.80.1.3:1883 check
>
> It works well as a cluster of mqtt brokers.
> Thank you very much again.
>
> Wang
> On 2015年09月02日 00:26, Tim Bain wrote:
>
>> In that case, you'll want each broker connected si
Problems like these are very difficult for third parties to debug since
they can be caused by many things, some of which are environmental and have
nothing to do with ActiveMQ (network dropouts, firewall issues, etc.). And
then on top of that, you're asking about a version of ActiveMQ that's three
James,
A full description of the TCP protocol is out of scope for this mailing
list, though I highly encourage you to pick up a book about the subject
(sorry, I don't have any specific recommendations, though I'm sure the
Internet does) or read up on it at one of the various websites on the
subjec
For me, the weirdest part about this is the fact that a full GC somehow
fixes is. GCs aren't supposed to do anything that would affect the
application (garbage should always be dead, so there should be no effect to
anything that's still alive), so that's really strange.
If you start ActiveMQ with
ng to JRE7 might
> solve the problem.
>
> On 3 September 2015 at 17:57, Tim Bain wrote:
>
> > For me, the weirdest part about this is the fact that a full GC somehow
> > fixes is. GCs aren't supposed to do anything that would affect the
> > application (garbage sh
Can you provide a link to the documentation you're referring to?
On Sep 10, 2015 3:52 PM, "mtod" wrote:
> Thanks for the quick reply.
>
> Good catch that resolved the problem.
>
> I was using the default activemq config and I guess they have not updated
> the documentation.
>
> Mike
>
>
>
>
> --
The interesting line in the stack trace is
org.apache.activemq.store.jdbc.JDBCPersistenceAdapter.getLastMessageBrokerSequenceId().
(The 5.11.1 version of that code is at
http://grepcode.com/file/repo1.maven.org/maven2/org.apache.activemq/activemq-jdbc-store/5.11.1/org/apache/activemq/store/jdbc/JDB
I believe that setting prefetch=0 only changes how that message gets
dispatched to the first consumer (pull vs. push); once it's there, I
wouldn't expect the broker's behavior w.r.t. that dispatched message to
change (and indeed you say it didn't). But as you say, closing the
connection returns al
If there are multiple consumers, having each consumer reject out-of-order
messages as Art suggested is only going to work if you use message groups,
so take his last paragraph as good general advice but ignore it if you
choose to implement the rest of his suggestions.
But if you're using message g
Can you just set the JMSExpiration message attribute (
http://activemq.apache.org/activemq-message-properties.html) so you know
you don't have any pending messages older than a certain interval? Then
the broker will automatically age off old messages, without you having to
write any code.
On Sep 1
It looks like the connection was closed because the inactivity monitor
didn't receive any data (keep-alive or otherwise) for longer than its
timeout. The keep-alive packets are supposed to be sent every 1/3 of the
keep-alive interval if no other content is being sent.
That can happen if there is
In that situation, broker 2 will PFC broker 1, and then (later, once broker
1 hits a usage limit) broker 1 will PFC the producer.
In a NoB, each broker acts as a producer or a consumer on each
networkConnector, so the brokers experience the same interactions that
clients do, and any interaction be
orked that way, but I wasn't sure. Thank you.
>
> Is there a specific Test Case that proves this?
>
> From: Tim Bain [via ActiveMQ] [mailto:
> ml-node+s2283324n4702052...@n4.nabble.com]
> Sent: Wednesday, September 16, 2015 11:04 PM
> To: Billy Buzzard
> Subject: Re: Do
As described at the bottom of
http://activemq.apache.org/how-do-durable-queues-and-topics-work.html, if
you want durable subscriptions to work, you can't allow the client ID to
change following a restart. Set it yourself, explicitly, to guarantee it
doesn't change.
On Sep 17, 2015 11:49 AM, "tras"
above) and it halts the Producer Flow from Broker B, but I'm not
> seeing anything that indicates the flow can be resumed.
>
> Can you suggest some ways to isolate this problem? I was thinking of
> using the existing Producer Flow Control test and modifying the test to
&
Have you looked in JConsole to see if there are any actions exposed via JMX
that would allow you to remove that subscription? I've never run 5.7.0 so
I can't say whether anything is available in that version (or any version,
for that matter), but it's where I'd look.
On Sep 18, 2015 7:01 AM, "tras
In your scenario, are both consumers already connected when the messages
are sent?
On Sep 18, 2015 4:59 PM, "sburczymucha" wrote:
> Hello,
>
> I would like to know if I understand how messages are processed in
> following
> configuration. Supposed I have a connection with message prefetch = 5, an
issue.
Tim
On Fri, Sep 18, 2015 at 7:04 PM, Tim Bain wrote:
> Have you looked in JConsole to see if there are any actions exposed via
> JMX that would allow you to remove that subscription? I've never run 5.7.0
> so I can't say whether anything is available in that version (
OK, so I'm assuming then that your actual question relates to the situation
where both consumers are already connected when the first message is sent,
since you just demonstrated that you know what happens when messages are
sent when only one consumer is connected and since the same thing happens
w
If you "force" a full GC (as much as it's possible to do that since it's
really just a suggestion), does memory usage drop substantially? If not,
it's not the garbage collector's fault, you're using all of your memory,
either because there's a memory leak (unlikely given how many other people
run
Also, I see you're using a DeserializerCache, which strongly implies
there'll be some caching going on; are you sure you've configured the
caching behavior appropriately for your JVM size and expected use case?
On Sep 21, 2015 8:22 PM, "Tim Bain" wrote:
> If you &
This is the first search result when I Google for 'activemq advisory':
http://activemq.apache.org/advisory-message.html
If you're not trying to explicitly use advisory topics yourself, you can
ignore them; brokers in a network of brokers use them to figure out where
consumers are, but you don't ha
Are there any messages on the queues in your broker? Lots of
unconsumed/unacknowledged messages would certainly explain what you're
seeing, and with PFC disabled there would be nothing to prevent the broker
from blowing right through the memory limit (till it ran into an actual
memory limit when i
Are there any topics with unconsumed messages for offline durable
subscribers? Queues aren't the only place messages can stick around.
You might get more information by using information exposed via JMX; it
might help you find where there are still messages if indeed there are
some.
On Sep 22, 20
That can happen if you have topics whose consumers don't consume all the
messages, either due to selectors or due to consumers being offline when
messages are published. (Do you have topics? And are advisory messages
enabled?) If neither of those are the case, there are probably unconsumed
messa
ssue
> and one of the "theories" is that messages sent to Advisory topics is not
> being freed because we see [sometimes very large] numbers of messages
> enqueued, no messages dequeued, and a non-zero number of consumers. Is
> that something that's possible?
>
>
>
Your efforts are focused on the wrong JVM. The broker's fine, and your
webapp is not, as shown by the two screenshots you shared.
On the broker side, what you're showing looks like perfectly normal JVM
garbage collection, though it looks like Young Gen is only 3/8ths of the
full heap, which is lo
I was hoping that one of the devs (Gary, Tim, Art, etc.) would jump into
this discussion, but since they didn't, please submit a JIRA bug to bring
this issue to their attention. If the problem is that the OpenWire
versions aren't backwards-compatible, that's either a bug (I was under the
impressio
I'm having trouble figuring out why AMQ1 and AMQW1 both have a
networkConnector that points to the same endpoint. Can you explain how
10.0.1.53:61616 and 10.0.1.200:61616 map to AMQ1/AMQ2/AMQW1/AMQW2?
Tim
On Sun, Sep 13, 2015 at 6:56 PM, atchijov wrote:
> I am trying to make ActiveMQ 5.12 and
Try your original configuration with duplex="true" on all networkConnectors.
On Sep 23, 2015 8:15 AM, "atchijov" wrote:
> Hi Tim,
> Sorry for not been clear enough.
>
> AMQ1 has IP 10.0.1.53 and its configuration points to 10.0.1.200:61616
> AMQ2 has IP 10.0.1.200 and its configuration poin
Sorry, not all, only the AMQ-AMQW ones. The AMQ1-AMQ2 connection is
already connected in both directions.
On Sep 23, 2015 8:31 AM, "Tim Bain" wrote:
> Try your original configuration with duplex="true" on all
> networkConnectors.
> On Sep 23, 2015 8:15 AM, &
Excellent, sounds like a valuable learning experience, which is one of the
things that internships are all about.
On Sep 23, 2015 2:20 PM, "mhempleman" wrote:
> Thanks for your help Tim. It was the application heap size. I upped it to
> 2G and the errors disappeared. I think I didn't see the e
In a master/slave cluster, only one broker is active at a time; otherwise
you have a network of brokers, not master/slave. Does that describe your
intended configuration?
If so, what you're seeing indicates that both brokers are active
simultaneously. Try to confirm that by hitting both brokers'
What about the subscribers on your topics? (Does anyone subscribe directly
to the topic, or is consumption purely from the queues on the virtual
topics?) And are any selectors in use for any client?
I'm curious whether the KahaDB files are cleaned up 1) following a broker
restart, and 2) if you
Please tell me you meant 5.12, not 5.1.2.
On Sep 26, 2015 5:08 PM, "mhempleman" wrote:
> Jboss 6.1 EAP
> ActiveMQ 5.1.2
>
> I'm trying to connect Jboss to an ActiveMQ high availability cluster. When
> I kill one ActiveMQ server and Jboss fails over to the backup server
> everything works as expe
Is the producer your last non-daemon thread? If so, the JVM will exit when
that thread exits, but you could use a synchronization construct such as a
countdown latch or a call to Thread.join() to make the producer thread not
exit till the consumer does.
On Sep 28, 2015 7:21 PM, "mfan" wrote:
> I
es
> | org.apache.activemq.broker.region.Queue | RMI TCP
> Connection(8)-130.20.132.42
> 2015-09-28 19:09:04,669 | INFO | queue://Queue-1438296198618 purged of 0
> messages | org.apache.activemq.broker.region.Queue | RMI TCP
> Connection(8)-130.20.132.42
>
> On Mon, Sep 28, 2015 at 7:05
They're mutually exclusive within a specific cluster of nodes, but you can
compose a network of brokers out of multiple master/slave cluster.
So (to pick a random example) if you have clusters made of M1/S1, M2/S2,
and M3/S3 and wanted a NOB with connections between clusters 1 and 2 and
between 2
I trust you've confirmed that there are no messages pending on topics as
well as queues? JMX is the easiest way I know to confirm that if you
haven't already done so.
Are some but not all journal files being deleted, or none at all?
On Sep 28, 2015 8:44 AM, wrote:
> Hello,
>
> How is it that ou
ve within the same
> master/slave group, but the same does not hold true for "cluster" given its
> ambiguity.
>
> Raffi
>
> -Original Message-
> From: tbai...@gmail.com [mailto:tbai...@gmail.com] On Behalf Of Tim Bain
> Sent: Tuesday, September 29, 2015 12:5
Most people on both Windows and Linux use the Java or C++ client libraries
to read and write messages from/to ActiveMQ. Do you have requirements that
preclude that approach? If so, what are you trying to do and under what
constraints, so we can try to help you find a solution that works for you?
; onMessage() will be invoked every time when the message comes in. Without
> knowing the number of messages beforehand, do I know when it will be
> finished ? Did I think wrong on what you have suggested ?
>
> Thanks.
>
> On Mon, Sep 28, 2015 at 9:46 PM, Tim Bain wrote:
>
ad a pending TX
> or something.
> >
> >
> > Also: is there any pending XA Transactions?
> >
> > On Tue, Sep 29, 2015 at 9:29 AM, wrote:
> >
> > > All Topic queue sizes in JConsole show 0, and dispatch counts show 0.
> > Any
> > > other
f journal files, you probably don't have
zero unconsumed messages.
Tim
On Wed, Sep 30, 2015 at 8:42 AM, Tim Bain wrote:
> If anything, you want smaller files not bigger ones, because you're hoping
> for a situation where for a given file, all of the messages in the file
> have thei
Given that the code in question is in DefaultMessageListenerContainer
(Spring), I'd encourage you to post to their mailing list and/or submit a
bug report to them and see if maybe there's something for them to fix.
Setting the timeout sounds like a reasonable workaround, but it would be
good to hav
Unfortunately I don't know of any tools for inspecting KahaDB files, though
someone else here might. Sorry, I've never done much with KahaDB, so what
I know comes from the wiki page and from mailing list posts like this one.
Tim
On Tue, Sep 29, 2015 at 3:33 PM, fmansoor wrote:
> Thanks Tim,
>
That's all true when the consumer is connected or when a network of brokers
is statically configured. If no consumer is currently connected to a
dynamically configured NOB, messages will stay on the broker to which
they're first published until a consumer connects, at which point they'll
be forwar
Both pages appear to have fallen victim to a backwards incompatibility
between the old version of the Include Page macro and the newer version of
Confluence. I've fixed both pages by simply deleting the macro on each
page and recreating it (which should fix this issue on any future pages on
which
Your assumption is correct, networks of brokers allow messages to be
forwarded across the network to be consumed by consumers on brokers other
than the one on which the message was produced.
Have you disabled advisory messages and/or explicitly included/excluded any
destinations? And have you don
A network of brokers can be used for HA as long as the network topology
will remain fully connected in the face of N failures (for whatever value
of N you choose to support, probably 1) and the clients' failover URIs will
always contain a live broker in the face of the same failure(s).
But you cou
BTW, I'm not convinced that a different network topology will avoid the
exception you first asked about. Someone who knows XA transactions (which
is not me, sorry) needs to look at that.
On Oct 1, 2015 7:06 AM, "Tim Bain" wrote:
> A network of brokers can be used for HA as l
ards,
>
> Barry
>
>
> -Original Message-
> From: tbai...@gmail.com [mailto:tbai...@gmail.com] On Behalf Of Tim Bain
> Sent: Thursday, October 01, 2015 8:56 AM
> To: ActiveMQ Users
> Subject: RE: Network of Brokers - Putting/Getting from a queue
>
> Your assumption is co
a dynamic queue called TEST?
>
> Regards,
>
> Barry
>
>
> -Original Message-
> From: tbai...@gmail.com [mailto:tbai...@gmail.com] On Behalf Of Tim Bain
> Sent: Thursday, October 01, 2015 9:13 AM
> To: ActiveMQ Users
> Subject: RE: Network of Brokers - Putting/Getting
Great, thank you!
On Oct 1, 2015 11:01 AM, "gijsbert802" wrote:
> Thanks Tim.
>
> I created an issue: https://issues.apache.org/jira/browse/AMQ-5995
>
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/ActiveMQ-exits-on-startup-with-UTFDataFormatException-bad-strin
You might be able to use Camel routes embedded in the broker to move
messages from each topic into a single new queue. That would preserve
message ordering even when the consumer is disconnected.
Note that I've never used embedded Camel routes, so I can't say for sure
that the approach would work
e on BrokerA?
>
> Regards,
>
> Barry
>
>
> -Original Message-
> From: tbai...@gmail.com [mailto:tbai...@gmail.com] On Behalf Of Tim Bain
> Sent: Thursday, October 01, 2015 10:20 AM
> To: ActiveMQ Users
> Subject: RE: Network of Brokers - Putting/Getting from
I don't know for sure, but I would expect that it would not be possible to
do that, since both endpoints are hosted by Jetty and it's Jetty that's
handling the authentication.
On Sep 28, 2015 3:50 AM, "Aleksandar Ivanisevic"
wrote:
> Hi,
>
> Is it possible to have different users for the console
om Broker A is not forwarded to it.
>
> Regards,
>
> Barry
>
> -Original Message-
> From: tbai...@gmail.com [mailto:tbai...@gmail.com] On Behalf Of Tim Bain
> Sent: Friday, October 02, 2015 1:21 AM
> To: ActiveMQ Users
> Subject: RE: Network of Brokers - Putting/Gettin
Can someone provide a list of which transports are allowed in
broker-to-broker networkConnectors and which are not?
Obviously tcp works. Barry Barnett's recent thread made it clear that ssl
works (though maybe not for duplex connections), which was news to me.
http://activemq.apache.org/networks-
> Regards,
> >
> > Barry
> >
> >
> > -Original Message-
> > From: tbai...@gmail.com [mailto:tbai...@gmail.com] On Behalf Of Tim Bain
> > Sent: Monday, October 05, 2015 9:45 AM
> > To: ActiveMQ Users
> > Subject: Transports availab
y, October 05, 2015 10:12 AM
> To: users@activemq.apache.org
> Subject: Re: Transports available for networkConnectors [ EXTERNAL ]
> Importance: High
>
> The auto transport should work but only if the activemq-broker jar is on
> the classpath. It would just use OpenWire.
>
&g
Do you have control over the consumer? Can you force it to disconnect
(e.g. by stopping the client process or by attaching a debugger and pausing
all threads) while you examine the message?
On Mon, Oct 5, 2015 at 2:02 PM, thenk24 wrote:
> Is there any way to put inhibit a queue so that a consum
What's the question here? Why temp usage exceeds TempLimit? Why
StoreUsage for the three brokers seems unrelated to disk usage? Why three
brokers in a LevelDB cluster have vastly different StorePercentUsage
values? Something else?
Do you have Producer Flow Control (
http://activemq.apache.org/
Have you tried using an XML config file instead of the broker URI syntax?
I can't help with the latter (no prior experience with broker URIs, sorry),
but lots of people on this list have experience getting an XML config file
to work, so you'd have a better shot at getting help.
Tim
On Mon, Oct 5,
You may need to set maxReconnectAttempts and startupMaxReconnectAttempts to
get control back from the failover transport when no brokers are
available. The failover transport's documentation (
http://activemq.apache.org/failover-transport-reference.html) explicitly
says that the timeout URI option
Pending counts are preserved because you have persistent messages and a
persistence store and there are messages in it when the broker started, not
because the broker actually saved the count. As Chris said, there's
currently no means to preserve that information across a restart, though
you could
TCP can be tuned to deal better with bursty traffic if you prevent the
congestion window from closing on an idle connection via the
tcp_slow_start_after_idle kernel setting. But if I remember correctly, you
have to control the settings of both hosts or it won't do much good. And
nothing says you'
If you're hoping for help from anyone other than the OP, you'll want to
provide the same information about your clients that Art asked the OP for.
On Oct 8, 2015 8:35 AM, "dhananjay.patkar"
wrote:
> We are experiencing same issue, with activemq 5.10 broker and consumers are
> created through came
To clarify, I mean that you should post the actual Camel routes you're
using.
On Oct 8, 2015 1:48 PM, "Tim Bain" wrote:
> If you're hoping for help from anyone other than the OP, you'll want to
> provide the same information about your clients that Art asked the O
Do you have a procedure that will reliably reproduce both problems you
describe (negative consumer counts and non-delivery of messages) each time
you run it?
Have you used a JMX viewer such as JConsole to examine the subscriptions on
each destination both when the broker is in a good state and whe
When you restarted the broker after reconfiguring it, did you remember to
set the environment variable to the new encryption password?
On Oct 8, 2015 8:36 AM, "rosy.sal...@murex.com"
wrote:
> I'm using activemq 5.11.1. I applied password encryption using the
> following
> url: http://activemq.apa
If you want redelivery to other consumers rather than just the one to which
the message was first delivered (which is how I interpreted your paragraph
about message-driven beans) and you can live with out-of-order delivery,
I'd think you'd want to configure broker redelivery as described at the
bot
I believe that the discarded message count in JMX is incremented (and so
I'd expect an advisory message to be generated, though I've never looked
for one to confirm that) when a message expires while still on the broker.
(I believe I've observed that if the message expires after being dispatched
to
I've never used Derby so I could be wrong, but I'd guess you just need the
JAR available at http://db.apache.org/derby/derby_downloads.html on the
classpath.
Tim
On Oct 12, 2015 12:30 PM, "J_S" wrote:
> I'm new to ActiveMQ.
>
> I'm trying to set up a Master/Slave configuration of ActiveMQ using
Happy to help, and I'm glad you got it working.
On Oct 12, 2015 2:31 PM, "J_S" wrote:
> Thank you very mutch, Tim!
> It works now :)
> You helped me a lot as a new to java.
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/Problem-using-derby-jdbc-EmbeddedDataSour
You should post your consumer code and your broker config. It's possible
to set settings in both of them that would result in the behavior you
describe, so I want to make sure everything is OK there.
Can you confirm that you see a connection from the embedded broker to the
standalone broker when
Boris,
That's an old enough bug that I think we're better off creating a new bug
in JIRA and simply linking the two together. Can you please submit it,
along with any configuration or procedures you needed to reproduce the
problem reliably?
Tim
On Oct 9, 2015 8:15 AM, "boriss" wrote:
> I'm us
umer try again. If you do that then be sure to
add some message property to keep count of how many times it's failed, so
the task doesn't keep sending it.
Andy.
agumbre...@tomitribe.com
http://www.tomitribe.com
Sent from my mobile device.
- Reply message -
From: "Tim Bain&
You want help debugging a memory leak in a version that was released 6
years ago (today, coincidentally)? Maybe someone will take pity on you and
help you out, but I'd be surprised. Upgrade to a modern version if you
want the best chance of getting someone to help you.
On Oct 13, 2015 8:39 AM, "c
So you're using ActiveMQ, Camel, and Spring, and the problem (whether it's
your error or a bug) could be in any of them. I've got to say, given that
this isn't clearly a problem with any one of those technologies, I'd post
this to StackOverflow and tag it with all three technologies and hope that
I'd say the tried and true approach to handling out-of-space exceptions is
to prevent them from happening in the first place: set the storeLimit and
tempLimit to appropriately large values, enable PFC to ensure you don't
exceed them, and host them in a storage location that you can guarantee
will a
Do your three clients use different clientIDs?
Also, it sounds like you're using durable subscriptions; is that accurate?
On Oct 14, 2015 11:44 AM, "wuwufen" <1767549...@qq.com> wrote:
> oh thanks
>
>
>
> --
> View this message in context:
> http://activemq.2283324.n4.nabble.com/use-mqtt-receive-
How many connections are in your connection pool?
On Oct 14, 2015 11:47 AM, "kal2" wrote:
> The FutureResponse becomes the bottleneck when we use more than 10 threads
> to
> send msgs out through connections.. We need to support large number of
> threads on producer side and they also send jms ms
Um, that looks like Producer Flow Control kicking in, which would result in
warnings in the broker's logs. Is that what you're seeing?
On Oct 14, 2015 11:47 AM, "dhananjay.patkar"
wrote:
> I was able to capture a thread dump on the jvm on which consumers are
> vanishing.
> The threads are blocke
Are you able to reliably reproduce this in a standalone test
configuration? If so, does the same behavior happen if you change to a
dynamic network? (That is, does your staticBridge setting actually
contribute to the behavior?)
Tim
On Oct 14, 2015 1:22 PM, "fedemoya" wrote:
> I'm running two a
Either Nabble oe the mailing list itself appears to have not acknowledged
either of my two previous replies from 36 hours ago in this thread, since
they're not on the Nabble web page. That's annoying, especially since I
wasn't given any notification that anything was wrong so I could fix it.
Here
MOTE_BROKER_URL +
")" when creating your ActiveMQConnectionFactory and see if the behavior
changes?
Also, your current setup doesn't provide fault tolerance for the standalone
broker itself. Do you need it to?
Tim
On Oct 15, 2015 6:52 AM, "Tim Bain" wrote:
> Eithe
401 - 500 of 2340 matches
Mail list logo