activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jim Gomes" <e.se...@gmail.com>
Subject Re: thread deadlock issue in NMS
Date Wed, 29 Oct 2008 17:31:22 GMT
Thanks for following up on is.  If you are able to narrow this down,
it would be much appreciated.  I know threading issues can be hard to
reproduce, especially when other things take your attention away.

I'll try and take a look again at some of the threading code to see if
I can find any obvious issues.

-Jim

On 10/29/08, mkeenan <keenan@p2sol.com> wrote:
>
>
> semog wrote:
>>
>> Thanks.  A helpful bit of information is a comparison of NMS 1.1 to NMS
>> 1.0,
>> since there was some major re-work done on the threading code between the
>> two versions.  Would you be able to test these different configurations?
>>
>
> I don't think I can do a apples-to-apples comparison because 1.0 was not
> thread safe due to the sessions collection not being synchronized.  To work
> around this issue, back when we used the NMS 1.0 library we used a separate
> connection/session for each consumer.  Although this is a slight abuse of
> resources we have a low number of consumers and it was working fine.
>
> When we switched to NMS 1.1 (primary motivation was the thread safety) I
> altered the configuration to share the connection. For testing, I might
> switch it back to one-connection-per-consumer and see if that changes the
> behavior we're observing.
>
> It might be a week or two before I can spend significant time on this due to
> other priorities.
> --
> View this message in context:
> http://www.nabble.com/thread-deadlock-issue-in-NMS-tp20180687p20227395.html
> Sent from the ActiveMQ - User mailing list archive at Nabble.com.
>
>

Mime
View raw message