Note that you still should get an error if the send fails in async mode - its just that you'll be notified in the ExceptionListener rather than receiving an exception in the calling thread (which makes it a bit hard to associate the message to the exception - I guess we really should make sure that the exceptions we throw also have the Message easily available).
On 4/24/06, Rob Davies <[EMAIL PROTECTED]> wrote: > Hi Reza, > > that's correct - it's a trade-off between performance and reliability > - so in ActiveMQ 4.0 we opted for safety first. > > cheers, > > Rob > > On 24 Apr 2006, at 08:17, reza aliakbary wrote: > > > Hi Rob, > > > > Yes, You are right. I changed the property and now the benchmark > > performs quite fast with the snapshot version. But what would I > > loose if I set useAsyncSend = true? I think I get no exception if > > a message cannot be delivered, right? > > > > Cheers, > > Reza > > > > Rob Davies <[EMAIL PROTECTED]> wrote: Hi Reza, > > > > this is good feedback - something that has changed between activemq > > 3.x and 4.0 is that by default messages sent from a MessageProducer > > now wait for a receipt from the message broker that the message has > > been received. So a property you could set on your Connection Factory > > is to set useAsyncSend = true. > > > > I'd be interested to know if this makes a difference for you. > > > > cheers, > > > > Rob > > > > On 23 Apr 2006, at 15:25, reza aliakbary wrote: > > > >> Dear Rob, The current snapshot performs better than RC2 but still > >> it is too slower than 3.2.2 . With a benchmark in the same > >> environment I got the following times: > >> 3.2.2 ----> 8 sec. > >> SNAPSHOT ----> 43 sec. > >> RC2 --- > 77 sec. > >> > >> So you see the diffrence is a lot(8 vs 43). > >> > >> Thanks, > >> Reza > >> > >> Rob Davies wrote: could you try the latest > >> snapshot: http://cvs.apache.org/repository/ > >> incubator-activemq/distributions/ - I've only started ActiveMQ 4.0 > >> has only started in the past week :) > >> > >> On 23 Apr 2006, at 08:06, reza aliakbary wrote: > >> > >>> Dear James, I also tested with RC2 but it is still slow. We are > >>> interested of the features you provided in version 4.0 but we > >>> don't want to loose performance. I hope you would reach to the > >>> performance of 3.2.2 in next releases of 4.0 . > >>> > >>> Cheers, > >>> Reza > >>> > >>> > >>> James Strachan wrote: Which version of > >>> 4.0 are you testing? FWIW we've only recently started > >>> tuning 4.x heavily, certainly SVN HEAD is looking quite good. > >>> > >>> > >>> On 4/22/06, reza aliakbary wrote: > >>>> Hello, > >>>> > >>>> I used a simple benchmark to compare the performance of 3.2.2 vs > >>>> 4.0 . I > >>>> suprised why 3.2.2 performs very better than 4.0, I thought I > >>>> could be wrong > >>>> or maybe my configuration has a problem but I couldn't find > >>>> anything bad in > >>>> configurations(I used default and without persistency). I have > >>>> attached > >>>> codes that I used. Please let me know why 3.2.2 performs better > >>>> than 4.0. > >>>> The benchmark works with a queue asynchronously. > >>>> Run the consumer to register consumers and them run the producer. > >>>> > >>>> Best Regards, > >>>> Reza > >>>> > >>>> > >>>> __________________________________________________ > >>>> Do You Yahoo!? > >>>> Tired of spam? Yahoo! Mail has the best spam protection around > >>>> http://mail.yahoo.com > >>>> > >>> > >>> > >>> -- > >>> > >>> James > >>> ------- > >>> http://radio.weblogs.com/0112098/ > >>> > >>> > >>> > >>> --------------------------------- > >>> Blab-away for as little as 1ยข/min. Make PC-to-Phone Calls using > >>> Yahoo! Messenger with Voice. > >> > >> > >> > >> > >> --------------------------------- > >> How low will we go? Check out Yahoo! Messenger's low PC-to-Phone > >> call rates. > > > > > > > > > > --------------------------------- > > How low will we go? Check out Yahoo! Messenger's low PC-to-Phone > > call rates. > > -- James ------- http://radio.weblogs.com/0112098/
