yes, 120 milliseconds, I wrote in a previous email: 1. Serialize the request 2. Send the request over to the other server 3. Deserialize and recreate the request (including invoke any listeners on the session) on the other server 4. Send a ACK message back to the original server 5. Receive the ACK message
also, make sure that your QA and your production servers use different multicast addresses. (property mcastAddress in server.xml) otherwise you are swapping sessions between QA and production Filip -----Original Message----- From: Randy [mailto:[EMAIL PROTECTED] Sent: Friday, May 21, 2004 4:45 PM To: 'Tomcat Users List' Subject: RE: Quiestion about Average cluster serialize/send So here is what I understand, and I want to make sure that I am reading this correctly The log entry says: Average cluster serialize/send time=120 ms for 300 requests For 300 requests it took on average 120 to serialize this data. (does this include broadcasting it to the other servers?) What kind of profile? thanks -----Original Message----- From: Filip Hanik (lists) [mailto:[EMAIL PROTECTED] Sent: Friday, May 21, 2004 4:23 PM To: Tomcat Users List Subject: RE: Quiestion about Average cluster serialize/send >Is it I am not understanding the message(oh what a shock that would be) >Box 1 >INFO: Average cluster serialize/send time=120 ms for 300 requests (36046ms). >Box 2 >INFO: Average cluster serialize/send time=8 ms for 300 requests (2401ms). I am not sure why we are seeing this. I couldn't tell more unless we actually ran a profile. Try to set your tcpSelectorTimeout to 10 instead of a 100, and let me know the results Filip -----Original Message----- From: Randy [mailto:[EMAIL PROTECTED] Sent: Friday, May 21, 2004 2:14 PM To: 'Tomcat Users List' Subject: RE: Quiestion about Average cluster serialize/send Sorry for the confusion. Originally I was In a test env ( "They are defend sized boxes" ) Where I was not replicating the same context's on both boxes. Once I got it working in a test evn(thanks for your help) I turned It on in production last night(re: message below with config info) The load balancing is kind of distributing the requests equally, but not really. There are some issues with aol proxies, so the balance is not totally equal. But they are usually pretty close. Is it I am not understanding the message(oh what a shock that would be) Box 1 INFO: Average cluster serialize/send time=120 ms for 300 requests (36046ms). Box 2 INFO: Average cluster serialize/send time=8 ms for 300 requests (2401ms). I am just looking at these two messages and thinking that I would have thought theys should be closer, assuming it is for the same amount of requests. I am thinking that there may be more to be looking at ?? It is very obvious I am way over my head, and I appreciate the help you guys are giving me Randy -----Original Message----- From: Filip Hanik (lists) [mailto:[EMAIL PROTECTED] Sent: Friday, May 21, 2004 1:50 PM To: Tomcat Users List Subject: RE: Quiestion about Average cluster serialize/send and your loadbalancing algorithm is distributing the requests equally. also, in a previous email you wrote >They are diffent sized boxes what does that mean? Filip -----Original Message----- From: Randy [mailto:[EMAIL PROTECTED] Sent: Friday, May 21, 2004 1:27 PM To: 'Tomcat Users List' Subject: RE: Quiestion about Average cluster serialize/send Thanks for any help These are two identical machines Rh9 Apache2 Tomcat 5.0.24 with hotfix Connected by gigaBit ethernet nics and switches. Box 1 is using mod_jk Box 2 is using mod_jk2 Box 1 has the real docroot Box 2 has box1 docroot mounted nfs (from previous message) Box 1 INFO: Average cluster serialize/send time=120 ms for 300 requests (36046ms). Box 2 INFO: Average cluster serialize/send time=8 ms for 300 requests (2401ms). As an FYI these are real servers in an LVS env. ============================================================================ ======== Both Boxes have same Cluster entries (different tcpListenAddress) <Cluster className="org.apache.catalina.cluster.tcp.SimpleTcpCluster" managerClassName="org.apache.catalina.cluster.session.DeltaManager" expireSessionsOnShutdown="false" useDirtyFlag="true"> <Membership className="org.apache.catalina.cluster.mcast.McastService" mcastAddr="228.0.0.4" mcastPort="45564" mcastFrequency="500" mcastDropTime="3000"/> <Receiver className="org.apache.catalina.cluster.tcp.ReplicationListener" tcpListenAddress="192.168.0.203" tcpListenPort="4001" tcpSelectorTimeout="100" tcpThreadCount="6"/> <Sender className="org.apache.catalina.cluster.tcp.ReplicationTransmitter" replicationMode="pooled"/> <Valve className="org.apache.catalina.cluster.tcp.ReplicationValve" filter=".*\.gif;.*\.js;.*\.jpg;.*\.htm;.*\.html;.*\.txt;.*\.pdf;"/> </Cluster> ============================================================================ ======== -----Original Message----- From: Filip Hanik (lists) [mailto:[EMAIL PROTECTED] Sent: Friday, May 21, 2004 12:46 PM To: Tomcat Users List Subject: RE: Quiestion about Average cluster serialize/send >Nr of bytes sent=369314 over 400 ==923 bytes/request average number of bytes sent over in each replication request >INFO: Average cluster serialize/send time=91 ms for 400 requests >(36640ms) the average time it takes to perform a replication request. IE, this is the overhead you have per request when using clustering. 91ms, this time includes: 1. Serialize the request 2. Send the request over to the other server 3. Deserialize and recreate the request (including invoke any listeners on the session) on the other server 4. Send a ACK message back to the original server 5. Receive the ACK message Usually what takes time is the deserialization of the request. Send us info on the two machines and the two configurations Filip -----Original Message----- From: Randy [mailto:[EMAIL PROTECTED] Sent: Friday, May 21, 2004 10:10 AM To: 'Tomcat Users List' Subject: Quiestion about Average cluster serialize/send Hello I got clustering up last night. In the log file I am seeing some thing that I was wondering if someone could help explain. What concerns me is the difference in Average cluster Is this the time it takes to serialize and send over the session info? I am trying to understand where the problem is? Thanks ======================== Box 1 log file snippet ======================== May 21, 2004 1:17:19 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=245486 over 200 ==1227 bytes/request May 21, 2004 1:19:24 AM org.apache.catalina.cluster.tcp.ReplicationValve addClusterSendTime INFO: Average cluster serialize/send time=120 ms for 300 requests (36046ms). May 21, 2004 1:36:31 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=302688 over 300 ==1008 bytes/request May 21, 2004 1:50:09 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=369314 over 400 ==923 bytes/request May 21, 2004 2:13:33 AM org.apache.catalina.cluster.tcp.ReplicationValve addClusterSendTime INFO: Average cluster serialize/send time=91 ms for 400 requests (36640ms). ======================== Box 2 log file snippet ======================== May 21, 2004 1:39:17 AM org.apache.catalina.cluster.tcp.ReplicationValve addClusterSendTime INFO: Average cluster serialize/send time=9 ms for 100 requests (947ms). May 21, 2004 1:54:41 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=158650 over 200 ==793 bytes/request May 21, 2004 2:30:03 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=216767 over 300 ==722 bytes/request May 21, 2004 3:28:29 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=283745 over 400 ==709 bytes/request May 21, 2004 4:16:34 AM org.apache.catalina.cluster.tcp.ReplicationValve addClusterSendTime INFO: Average cluster serialize/send time=9 ms for 200 requests (1846ms). May 21, 2004 4:31:52 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=354565 over 500 ==709 bytes/request May 21, 2004 5:13:15 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=436113 over 600 ==726 bytes/request May 21, 2004 5:40:14 AM org.apache.catalina.cluster.tcp.ReplicationValve addClusterSendTime INFO: Average cluster serialize/send time=8 ms for 300 requests (2401ms). May 21, 2004 5:42:29 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=511204 over 700 ==730 bytes/request May 21, 2004 6:04:17 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=587590 over 800 ==734 bytes/request May 21, 2004 6:30:18 AM org.apache.catalina.cluster.tcp.ReplicationTransmitter addStats INFO: Nr of bytes sent=657453 over 900 ==730 bytes/request --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED] --- Incoming mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --- Outgoing mail is certified Virus Free. Checked by AVG anti-virus system (http://www.grisoft.com). Version: 6.0.676 / Virus Database: 438 - Release Date: 5/3/2004 --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
