Further update: These are all Selector Aware Queues. One VirtualTopic, and multiple Selector Aware queues underneath it. I noticed one of the big queue that is not dispatching messages to the consumers has lot of unmatched messages. I think these unmatched messages are suppressing the dispatching of matched messages due to mixing of priorities.
On Mon, Nov 30, 2015 at 11:36 AM, Takawale, Pankaj < pankaj.takaw...@dowjones.com> wrote: > One more observation: > Number of consumers being displayed on Admin UI are way more than actual > consumers. > Admin UI is showing incorrect? counts: > Pending Messages: 568073 > Number of consumers 408 (I'm expecting only 69 consumers), > Messages Enqueued 354845 > Messages Dequeued 45383 > > Thanks > Pankaj > > > On Mon, Nov 30, 2015 at 10:02 AM, Takawale, Pankaj < > pankaj.takaw...@dowjones.com> wrote: > >> Yes, my messages has mixed priorities. >> >> I've around 100 queues, 1500 connections, 2500 consumers, and 20 >> producers. Broker processes around 6 million messages per day. >> This problem is occurring with only couple of high-volume queues. All >> other queues are working fine. >> >> >> I drained 300K messages by writing curl script to browse the queue >> (whenever script successfully browses queue, consumers would receive >> messages). Now queue browsing is not working as I'm getting GATEWAY_TIMEOUT >> messages. So around 500K messages are stuck in queue. >> >> Since it's a production environment, I can not test with the latest >> version right away. >> >> Thanks >> Pankaj >> >> >> >> >> >> >> >> >> >> On Sat, Nov 28, 2015 at 12:19 AM, Tim Bain <tb...@alumni.duke.edu> wrote: >> >>> The bug you linked to was pretty specifically about messages with mixed >>> priorities. Does that describe your scenario? >>> On Nov 27, 2015 12:58 PM, "Timothy Bish" <tabish...@gmail.com> wrote: >>> >>> > >>> > Just set your store version to 6 via >>> BrokerService#setStoreOpenWireVersion >>> > >>> > On 11/27/2015 02:29 PM, Takawale, Pankaj wrote: >>> > > I tried that, but ActiveMQ 5.12 does not start with the 5.11 >>> messages in >>> > > database. >>> > > I got the same exception mentioned in >>> > > https://issues.apache.org/jira/browse/AMQ-5995 >>> > > >>> > > Thanks >>> > > Pankaj >>> > > >>> > > On Fri, Nov 27, 2015 at 2:08 PM, Timothy Bish <tabish...@gmail.com> >>> > wrote: >>> > > >>> > >> On 11/27/2015 01:31 PM, Takawale, Pankaj wrote: >>> > >>> Version: 5.11.1 >>> > >>> Persistence: JDBC (Postgresql) >>> > >>> >>> > >>> Consumers are using Selector expressions to select messages based >>> on >>> > >>> JMSXGroupID. >>> > >>> >>> > >>> When a queue receives burst of messages, sometimes all messages >>> gets >>> > >> stuck >>> > >>> in queue, and consumers would not receive any messages. >>> > >>> >>> > >>> But, when I click on the queue through Admin UI, consumer apps >>> receive >>> > >>> bunch of messages, and queue freezes again. >>> > >>> >>> > >>> Is this bug related to AMQ-5853 >>> > >>> <https://issues.apache.org/jira/browse/AMQ-5853> ? >>> > >>> >>> > >>> >>> > >>> Thanks, >>> > >>> Pankaj. >>> > >>> >>> > >> Quick way to find out is to run with the latest release and see what >>> > >> happens. >>> > >> >>> > >> -- >>> > >> Tim Bish >>> > >> Sr Software Engineer | RedHat Inc. >>> > >> tim.b...@redhat.com | www.redhat.com >>> > >> twitter: @tabish121 >>> > >> blog: http://timbish.blogspot.com/ >>> > >> >>> > >> >>> > >>> > >>> > -- >>> > Tim Bish >>> > Sr Software Engineer | RedHat Inc. >>> > tim.b...@redhat.com | www.redhat.com >>> > twitter: @tabish121 >>> > blog: http://timbish.blogspot.com/ >>> > >>> > >>> >> >> >
