Hi all, I am trying to do some tests to see how long it takes for Zookeeper to write various sizes of data on each znode. I am using PHP-API from https://github.com/andreiz/php-zookeeper. I have 2 questions: 1. How do I write 4000 bytes - 1MB of data on each znode? My php client starts acting up beyond 4000 bytes and I am not able to edit the client. I am using VIM on Xubuntu VM 32 bit. 2. I have a 3 node (N1, N2 (leader), N3) cluster set-up. When I perform a single write on N1 I see weird timings on the zookeeper log files. Let's say I do a write to N1 which is a follower. I get the following result on the N1 and N2's log files: N1----------------
2014-05-07 10:51:01,732 [myid:1] - DEBUG [QuorumPeer[myid=1]/0:0:0:0:0:0:0:0:2181:CommitProcessor@161] - Committing request:: sessionid:0x145be8a7f4b0041 type:create cxid:0x536a4857 zxid:0x100000196 txntype:1 reqpath:n/a 2014-05-07 10:51:01,733 [myid:1] - DEBUG [CommitProcessor:1:FinalRequestProcessor@88] - Processing request:: sessionid:0x145be8a7f4b0041 type:create cxid:0x536a4857 zxid:0x100000196 txntype:1 reqpath:/test--10 2014-05-07 10:51:01,734 [myid:1] - DEBUG [CommitProcessor:1:FinalRequestProcessor@160] - sessionid:0x145be8a7f4b0041 type:create cxid:0x536a4857 zxid:0x100000196 txntype:1 reqpath:/test--10 N2---------------- 2014-05-07 10:51:01,705 [myid:2] - DEBUG [ProcessThread(sid:2 cport:-1)::CommitProcessor@171] - Processing request:: sessionid:0x145be8a7f4b0041 type:create cxid:0x536a4857 zxid:0x100000196 txntype:1 reqpath:n/a 2014-05-07 10:51:01,705 [myid:2] - DEBUG [ProcessThread(sid:2 cport:-1)::Leader@768] - Proposing:: sessionid:0x145be8a7f4b0041 type:create cxid:0x536a4857 zxid:0x100000196 txntype:1 reqpath:n/a 2014-05-07 10:51:01,740 [myid:2] - DEBUG [LearnerHandler-/10.22.70.20:47250:CommitProcessor@161] - Committing request:: sessionid:0x145be8a7f4b0041 type:create cxid:0x536a4857 zxid:0x100000196 txntype:1 reqpath:n/a 2014-05-07 10:51:01,741 [myid:2] - DEBUG [CommitProcessor:2:FinalRequestProcessor@88] - Processing request:: sessionid:0x145be8a7f4b0041 type:create cxid:0x536a4857 zxid:0x100000196 txntype:1 reqpath:n/a How is the operation first logged in N2 before logging in N1? This does not happen all the time, the order of logging for a single operation is always different during various write operations. Anyone knows why this could be happening? My question regarding the timing is only for logging. I know all the operations in zookeeper are ordered so do not confuse it with that. Also, the timings on my zookeeper servers are synchronized. Let me know if you need any more information! Many thanks, Ganesh
