Single message unconsumed

[Dhruba Bandopadhyay <bd...@gmail.com>]

Hi,

Frequently during my tests I've noticed that they complete to 99%.
That is to say that frequently one single message remains unconsumed
on the broker queue.  The consumers seem fine and I don't see any
errors.

What are the usual reasons for this happening?  I am using Spring
2.5.2 and ActiveMQ 4.1.1.  This is quite unsettling as it is obviously
making the system less predictable and less robust so any help would
be appreciated.

Thanks.

Re: Single message unconsumed

[Dhruba Bandopadhyay <bd...@gmail.com>]

I've attached a grep of one of the consumer log files.  It seems that
the derby backend may have been the problem with regard to locking.
It seems that over a period of time under high number of messages
derby gets confused.

What alternatives do I have?

My attempt to use mysql as a backend with activem4.1.1 failed with the
unrecognized tag mapping error (which I posted on here with no
response).

My attempt to use mysql as a backend with activemq5.0.0 succeeded but
it tripled the production time and doubled total consumption time.
Also 5.0.0 (just like fmb-5.0.0.9) has the 'could not recover message
from data store' problem which makes it a non-option anyway.

Any recommendations?  I feel I'm at a dead end somewhat and would
really appreciate some help.

Thanks.

On Mon, Apr 7, 2008 at 9:19 AM, Dhruba Bandopadhyay <bd...@gmail.com> wrote:
> Hi,
>
>  Frequently during my tests I've noticed that they complete to 99%.
>  That is to say that frequently one single message remains unconsumed
>  on the broker queue.  The consumers seem fine and I don't see any
>  errors.
>
>  What are the usual reasons for this happening?  I am using Spring
>  2.5.2 and ActiveMQ 4.1.1.  This is quite unsettling as it is obviously
>  making the system less predictable and less robust so any help would
>  be appreciated.
>
>  Thanks.
>