Re: Most optimum receiver - Prefetch how?

2012-11-29 Thread Rajesh Khan
Thanks for clearing that up On Thu, Nov 29, 2012 at 7:53 AM, Gordon Sim wrote: > On 11/29/2012 02:48 PM, Rajesh Khan wrote: > >> so should i send a session acknowledgment after processing every ten >> messages or after every message for optimum performance ? >> > > Doing so every 10 messages wil

Re: Most optimum receiver - Prefetch how?

2012-11-29 Thread Gordon Sim
On 11/29/2012 02:48 PM, Rajesh Khan wrote: so should i send a session acknowledgment after processing every ten messages or after every message for optimum performance ? Doing so every 10 messages will improve the throughput as compared with doing so every message. -

Re: Most optimum receiver - Prefetch how?

2012-11-29 Thread Rajesh Khan
so should i send a session acknowledgment after processing every ten messages or after every message for optimum performance ? On Thu, Nov 29, 2012 at 7:28 AM, Gordon Sim wrote: > On 11/29/2012 01:30 PM, Rajesh Khan wrote: > >> Thank you for your email and making things a lot clear. However what

Re: Most optimum receiver - Prefetch how?

2012-11-29 Thread Gordon Sim
On 11/29/2012 01:30 PM, Rajesh Khan wrote: Thank you for your email and making things a lot clear. However what if I send an asynchronous acknowledgment like session.Acknowledge(false); will it be similar in performance to sending a parameter less acknowledgment. Yes, because the default value

Re: Does Qpid messaging API create a new thread for every new connection?

2012-11-29 Thread trivedi_ravi13
Sorry, I was testing it on 0.6 version and bug seems to fixed in 0.7, we tried on 0.10 version and setting the environment variable QPID_MAX_IOTHREADS worked for us. Thank you for your help..!! :) -- View this message in context: http://qpid.2158936.n2.nabble.com/Does-Qpid-messaging

Re: Most optimum receiver - Prefetch how?

2012-11-29 Thread Rajesh Khan
Thank you for your email and making things a lot clear. However what if I send an asynchronous acknowledgment like session.Acknowledge(false); will it be similar in performance to sending a parameter less acknowledgment. Which one will be better for clients that receive high frequency messages? On

RE: What if a a message is sent and then a connection is instantly closed after that ?

2012-11-29 Thread Jan Bares
> > It works fine but after sending several thousands messages I got very > strange error: > > > > javax.jms.IllegalStateException: Object AMQConnection: > > Host: prgesbclu > > Port: 5672 > > Virtual Host: test > > Client ID: clientid > > Active session count: 0 has been closed > > at > org.ap

Re: Does Qpid messaging API create a new thread for every new connection?

2012-11-29 Thread Gordon Sim
On 11/29/2012 11:57 AM, trivedi_ravi13 wrote: Thank you, I Tried as you suggested, setting the environment variable had no change in number of threads getting created when I look at the total number of threads in /proc//status file. I am creating 6 different connection objects and tryi

Re: What if a a message is sent and then a connection is instantly closed after that ?

2012-11-29 Thread Gordon Sim
On 11/29/2012 11:55 AM, Jan Bares wrote: Hi, does this recommendation also apply to Java JMS client? I have a connection open all the time and create session, send message, close session for each message to be sent. Opening and closing a *session* for each message is a lot less expensive th

Re: Does Qpid messaging API create a new thread for every new connection?

2012-11-29 Thread trivedi_ravi13
Thank you, I Tried as you suggested, setting the environment variable had no change in number of threads getting created when I look at the total number of threads in /proc//status file. I am creating 6 different connection objects and trying to open connection on 6 different qpid brokers.

RE: What if a a message is sent and then a connection is instantly closed after that ?

2012-11-29 Thread Jan Bares
Hi, does this recommendation also apply to Java JMS client? I have a connection open all the time and create session, send message, close session for each message to be sent. It works fine but after sending several thousands messages I got very strange error: javax.jms.IllegalStateException: O

Re: All cluster members die simultaneously with store state "dirty"

2012-11-29 Thread Gordon Sim
On 11/29/2012 08:40 AM, Pavel Moravec wrote: Hi, qpidd brokers stopped as they run on un-quorable cluster: Nov 23 23:14:09 server_name corosync[27510]: [CMAN ] quorum lost, blocking activity Nov 23 23:14:09 server_name corosync[27510]: [QUORUM] This node is within the non-primary componen

Re: What if a a message is sent and then a connection is instantly closed after that ?

2012-11-29 Thread Gordon Sim
On 11/29/2012 05:41 AM, Rajesh Khan wrote: I have something like this in my C++ code due to certain reasons.And wanted to know if messages are sent this way and its effect on the broker sender.send(message); connection__->close(); return true; If thousands of messages are sent ev

Re: About batch fetching - Prefetch in qpid

2012-11-29 Thread Gordon Sim
On 11/29/2012 01:47 AM, Adam Zedan wrote: I am trying to understand batch fetching with the following practical example.Currently My Qpid Broker has 70 items in it. I am trying to implement batch fetching however I am not having any success. Here is what I am doing in C# receiver. R

Re: Most optimum receiver - Prefetch how?

2012-11-29 Thread Gordon Sim
On 11/28/2012 09:16 PM, Rajesh Khan wrote: I have something like the following code in the receiver. I am currently dealing with 1000's of messages in perhaps less than 30 seconds. I read in the broker manual that by default only one message is extracted by the receiver from the broker however us

Re: Does Qpid messaging API create a new thread for every new connection?

2012-11-29 Thread Gordon Sim
On 11/28/2012 08:11 PM, trivedi_ravi13 wrote: I have observed that a new thread is being created every time a new connection to the qpid broker is opened. I am using messaging APIs. Is there any way to restrict or stop it from creating new threads? Let's say when we have large number of conn

Re: All cluster members die simultaneously with store state "dirty"

2012-11-29 Thread Pavel Moravec
Hi, qpidd brokers stopped as they run on un-quorable cluster: Nov 23 23:14:09 server_name corosync[27510]: [CMAN ] quorum lost, blocking activity Nov 23 23:14:09 server_name corosync[27510]: [QUORUM] This node is within the non-primary component and will NOT provide any services. Nov 23 23: