<p dir="ltr">This is weird, I would have expected problems with the last message, but not in the middle of the sequence (that's why I suggested sending only 1 message). Maybe we need an an even lower max_credits...</p>
<p dir="ltr">Merry Christmas to you, too!</p>
<p dir="ltr">Dan</p>
<div class="gmail_quote">On 21 Dec 2012 16:41, "Radim Vansa" <<a href="mailto:rvansa@redhat.com">rvansa@redhat.com</a>> wrote:<br type="attribution"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
Hi Dan,<br>
<br>
I have ran the test on 4 nodes in hyperion (just for the start to see how it will behave) but with 100 messages (1 message is nothing for a statistician) each 10MB and I see a weird behaviour - there are about 5-10 messages received in a fast succession and then the nothing is received for several seconds. I experience this behaviour for both 200k and 500k credits. Is this really how it should perform?<br>
<br>
Merry Christmas and tons of snow :)<br>
<br>
Radim<br>
<br>
<h1>☃</h1><br>
<br>
----- Original Message -----<br>
| From: "Dan Berindei" <<a href="mailto:dan.berindei@gmail.com">dan.berindei@gmail.com</a>><br>
| To: "infinispan -Dev List" <<a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a>><br>
| Sent: Tuesday, December 18, 2012 8:57:08 AM<br>
| Subject: Re: [infinispan-dev] MFC/UFC credits in default config<br>
|<br>
|<br>
| Hi Radim<br>
|<br>
| If you run the test with only 2 nodes and FC disabled, it's going to<br>
| perform even better. But then as you increase the number of nodes,<br>
| the speed with no FC will drop dramatically (when we didn't have<br>
| RSVP enabled, with only 3 nodes, it didn't manage to send 1 x 10MB<br>
| message in 10 minutes).<br>
|<br>
| Please run the tests with as many nodes as possible and just 1<br>
| message x 10MB. If 500k still performs better, create a JIRA to<br>
| change the default.<br>
|<br>
| Cheers<br>
| Dan<br>
|<br>
|<br>
|<br>
|<br>
|<br>
| On Mon, Dec 17, 2012 at 4:55 PM, Radim Vansa < <a href="mailto:rvansa@redhat.com">rvansa@redhat.com</a> ><br>
| wrote:<br>
|<br>
|<br>
| Sorry I haven't specified the amount, I am a stupido... my tests are<br>
| working with 500k credits.<br>
|<br>
| UUPerf (JGroups 3.2.4.Final-redhat-1) from one computer in perflab to<br>
| another, 2 threads (default), 1000x sends 10MB message (default<br>
| chunkSize = 10000 * our entry size is usually 1kB) executed 3x<br>
|<br>
| 200k: Average of 6.02 requests / sec (60.19MB / sec), 166.13 ms<br>
| /request (prot=UNICAST2)<br>
| Average of 5.61 requests / sec (56.09MB / sec), 178.30 ms /request<br>
| (prot=UNICAST2)<br>
| Average of 5.49 requests / sec (54.94MB / sec), 182.03 ms /request<br>
| (prot=UNICAST2)<br>
|<br>
| 500k: Average of 7.93 requests / sec (79.34MB / sec), 126.04 ms<br>
| /request (prot=UNICAST2)<br>
| Average of 8.18 requests / sec (81.82MB / sec), 122.23 ms /request<br>
| (prot=UNICAST2)<br>
| Average of 8.41 requests / sec (84.09MB / sec), 118.92 ms /request<br>
| (prot=UNICAST2)<br>
|<br>
| Can you also reproduce such results? I think that suggests that 500k<br>
| behaves really better.<br>
|<br>
| Radun<br>
|<br>
|<br>
|<br>
|<br>
| ----- Original Message -----<br>
| | From: "Dan Berindei" < <a href="mailto:dan.berindei@gmail.com">dan.berindei@gmail.com</a> ><br>
| | To: "infinispan -Dev List" < <a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a> ><br>
| | Sent: Monday, December 17, 2012 12:43:37 PM<br>
| | Subject: Re: [infinispan-dev] MFC/UFC credits in default config<br>
| |<br>
| |<br>
| |<br>
| |<br>
| |<br>
| | On Mon, Dec 17, 2012 at 1:28 PM, Bela Ban < <a href="mailto:bban@redhat.com">bban@redhat.com</a> ><br>
| | wrote:<br>
| |<br>
| |<br>
| | Dan reduced those values to 200K, IIRC it was for UUPerfwhich<br>
| | behaved<br>
| | best with 200K. Idon't know if this is still needed. Dan ?<br>
| |<br>
| |<br>
| |<br>
| |<br>
| | I haven't run UUPerf in a while...<br>
| |<br>
| |<br>
| |<br>
| |<br>
| | On 12/17/12 12:19 PM, Radim Vansa wrote:<br>
| | > Hi,<br>
| | ><br>
| | > recently I have synchronized our jgroups configuration with the<br>
| | > default one shipped with Infinispan<br>
| | > (core/src/main/resources/jgroups-(tcp|udp).xml) and it has shown<br>
| | > that 200k credits in UFC/MFC (I keep the two values in sync) is<br>
| | > not enough even for our smallest resilience test (killing one of<br>
| | > four nodes). The state transfer was often blocked when requesting<br>
| | > for more credits which resulted in not completing it within the<br>
| | > time limit.<br>
| | > Therefore, I'd like to suggest to increase the amount of credits<br>
| | > in<br>
| | > default configuration as well, because we simply cannot use the<br>
| | > lower setting and it's preferable to have the configurations as<br>
| | > close as possible. The only settings we need to keep different<br>
| | > are<br>
| | > thread pool sizes and addresses and ports.<br>
| | ><br>
| |<br>
| |<br>
| | What value would you like to use instead?<br>
| |<br>
| | Can you try UUPerf with 200k and your proposed configuration and<br>
| | compare the results?<br>
| |<br>
| | Cheers<br>
| | Dan<br>
| |<br>
| |<br>
|<br>
|<br>
| | _______________________________________________<br>
| | infinispan-dev mailing list<br>
| | <a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br>
| | <a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev" target="_blank">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br>
| _______________________________________________<br>
| infinispan-dev mailing list<br>
| <a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br>
| <a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev" target="_blank">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br>
|<br>
|<br>
| _______________________________________________<br>
| infinispan-dev mailing list<br>
| <a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br>
| <a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev" target="_blank">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a><br>
<br>
_______________________________________________<br>
infinispan-dev mailing list<br>
<a href="mailto:infinispan-dev@lists.jboss.org">infinispan-dev@lists.jboss.org</a><br>
<a href="https://lists.jboss.org/mailman/listinfo/infinispan-dev" target="_blank">https://lists.jboss.org/mailman/listinfo/infinispan-dev</a></blockquote></div>