Anybody ran sample smoke Tests for Curator recepies?

81 views
Skip to first unread message

The Other Guy

unread,
Feb 28, 2013, 7:50:29 PM2/28/13
to curato...@googlegroups.com
really new to zookeeper... We ran few smoke tests on 3 servers and getting +/- 300ms to acquire a lock from a client. 

did anybody run smoke test(s) for curator recipes (anything similar to this https://cwiki.apache.org/ZOOKEEPER/servicelatencyoverview.html)

thank you

Jordan Zimmerman

unread,
Feb 28, 2013, 7:52:04 PM2/28/13
to curato...@googlegroups.com
Are you testing with the testing servers or a real ZK cluster? If it's the testing servers I wouldn't be surprised. For real servers that seems very high.

-Jordan

--
You received this message because you are subscribed to the Google Groups "curator-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to curator-user...@googlegroups.com.
To post to this group, send email to curato...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msg/curator-users/-/l6yjoVjuAtkJ.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

The Other Guy

unread,
Feb 28, 2013, 7:53:18 PM2/28/13
to curato...@googlegroups.com
real cluster, yes it seems crazy high to me as well, that's exactly why I am asking....  ;-)

Jordan Zimmerman

unread,
Feb 28, 2013, 8:09:42 PM2/28/13
to curato...@googlegroups.com
I just ran the test below and got these times. This is from my local machine to the Netflix 5 instance cluster at AWS. The first timing makes sense as the connection has to get set up. As a check, I also did the test with a raw ZooKeeper handle just creating some nodes. The average round trip was ~100ms. 

Elapsed 0: 1623
Elapsed 1: 161
Elapsed 2: 160
Elapsed 3: 161
Elapsed 4: 161
Elapsed 5: 161
Elapsed 6: 161
Elapsed 7: 161
Elapsed 8: 161
Elapsed 9: 163

        CuratorFramework client = CuratorFrameworkFactory.newClient(..., new ExponentialBackoffRetry(100, 3));
        client.start();
        try
        {
            for ( int i = 0; i < 10; ++i )
            {
                long start = System.currentTimeMillis();
                InterProcessMutex lock = new InterProcessMutex(client, "/lock");
                Assert.assertTrue(lock.acquire(10, TimeUnit.SECONDS));
                System.out.println("Elapsed " + i + ": " + (System.currentTimeMillis() - start));
                lock.release();
            }
        }
        finally
        {
            client.close();
        }


To view this discussion on the web visit https://groups.google.com/d/msg/curator-users/-/9FN1QNhtI7QJ.

Brandon Beck

unread,
Feb 28, 2013, 8:44:42 PM2/28/13
to curato...@googlegroups.com
Jordan I believe your test really measured the time for you to reach AWS from your local machine.  I ran the same test on a server in AWS against our 3 instance cluster in AWS and got the following results.  For me the average round trip was in the 10's of milliseconds.

Elapsed 0: 1162
Elapsed 1: 10
Elapsed 2: 10
Elapsed 3: 9
Elapsed 4: 17
Elapsed 5: 8
Elapsed 6: 14
Elapsed 7: 8
Elapsed 8: 8
Elapsed 9: 23

Jordan Zimmerman

unread,
Feb 28, 2013, 8:48:32 PM2/28/13
to curato...@googlegroups.com
Thanks Brandon. So, czechjiri, you should probably do some network level debugging on your system. FYI - our production ZooKeeper instances are m2.4xlarge machines. Also, remember that the more ZK instances you have the _slower_ writes get.

-Jordan

The Other Guy

unread,
Feb 28, 2013, 9:03:25 PM2/28/13
to curato...@googlegroups.com
Thanks a lot for great answers! This is exactly what I needed.
Reply all
Reply to author
Forward
0 new messages