Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Zookeeper, mail # user - Exceeded deadline by 16ms


+
Deepak Jagtap 2012-05-16, 23:39
+
李赫元 2012-05-17, 01:30
+
Deepak Jagtap 2012-05-17, 18:36
+
李赫元 2012-05-18, 05:36
+
Patrick Hunt 2012-05-17, 22:23
Copy link to this message
-
Re: Exceeded deadline by 16ms
Deepak Jagtap 2012-05-17, 23:01
Thanks Patrick, that helped!

Best Regards,
Deepak

On Thu, May 17, 2012 at 3:23 PM, Patrick Hunt <[EMAIL PROTECTED]> wrote:

> The deadline exceeded messages are generated by select - the client
> will select(timeout). If the requested timeout is exceeded by more
> than 10ms then that message is printed. Often this will happen in a
> virtualized client environment (not the case here from what you are
> saying), or an overloaded OS, etc...
>
> You are running 12 client processes, where the client code expects a
> 10second session timeout (so heartbeat responses w/in 3.3 seconds)
> along three vms running three servers all on the same laptop, it's
> likely that you are overloaded.
>
> Patrick
>
> On Thu, May 17, 2012 at 11:36 AM, Deepak Jagtap <[EMAIL PROTECTED]>
> wrote:
> > Hi,
> >
> > Thanks for quick reply!
> > I am seeing sequence of "exceeded deadline" message then  connection
> timed
> > out error messages.
> > My code has some usleep() calls, when a sequence of zoo_get and zoo_set
> > returns ZBADVERSION error due to race condition,
> > it invokes usleep() for random time(between 5ms to 100ms) and then retry
> > the same operation. Does this may be cause any issues?
> >
> > Please find following log dump:
> >
> > 2012-05-17 11:10:12,910:6709(0x7f12d52d2700):ZOO_WARN@zookeeper_interest
> @1557:
> > Exceeded deadline by 14ms
> > 2012-05-17 11:10:13,200:6709(0x7f12d52d2700):ZOO_WARN@zookeeper_interest
> @1557:
> > Exceeded deadline by 11ms
> > .
> > .
> > 2012-05-17 11:10:13,270:6709(0x7f12d52d2700):ZOO_WARN@zookeeper_interest
> @1557:
> > Exceeded deadline by 11ms
> > 2012-05-17 11:10:13,745:6709(0x7f12d52d2700):ZOO_WARN@zookeeper_interest
> @1557:
> > Exceeded deadline by 11ms
> > 2012-05-17
> > 11:10:14,164:6709(0x7f12d52d2700):ZOO_ERROR@handle_socket_error_msg
> @1644:
> > Socket [192.168.1.236:2181] zk retcode=-7, errno=110(Connection timed
> out):
> > connection timed out (exceeded timeout by 0ms)
> > operation timeout
> > operation timeout
> > 2012-05-17 11:10:14,164:6709(0x7f12d52d2700):ZOO_INFO@check_events@1701:
> > initiated connection to server [192.168.1.234:2181]
> >
> >
> > Thanks & Regards,
> > Deepak
> >
> > On Wed, May 16, 2012 at 6:30 PM, 李赫元 <[EMAIL PROTECTED]> wrote:
> >
> >> Hello,
> >>
> >> Did you get the "ZOO_SESSION_EVENT" with "ZOO_EXPIRED_SESSION_STATE"
> >> on watch callback event or just simply debug information print on
> >> screen showing "Exceeded deadline by 16ms"?
> >>
> >> Thanks.
> >>
> >>
> >> 2012/5/17 Deepak Jagtap <[EMAIL PROTECTED]>:
> >> > Hi All,
> >> >
> >> > I am using Zookeeper 3.4.3 sever and my setup is as follows:
> >> >
> >> > I have a Quorum of 3 Zookeepers servers (installed on 3 virtual
> machines)
> >> > on my laptop.
> >> > Then I start 12 Zookeeper client processes on the same machine which
> >> start
> >> > updates on Zookeeper database.
> >> > After running test for a while, sometimes some of the client process
> got
> >> > disconnected from Zookeeper sever with following error:
> >> > *Exceeded deadline by 16ms* and then says session timeout.
> >> > Client connects with timeout value of 10000ms.
> >> >
> >> > Zookeeper server configuration is as follows:
> >> > tickTime=2000
> >> > initLimit=10
> >> > syncLimit=5
> >> > dataDir=/tmp/zookeeper
> >> > clientPort=2181
> >> > server.1=192.168.1.234:2888:3888
> >> > server.2=192.168.1.235:2888:3888
> >> > server.3=192.168.1.236:2888:3888
> >> >
> >> > Does anybody facing similar problem? Please let me know if there
> anythin
> >> > wrong wit configuration.
> >> >
> >> > Thanks & Regards,
> >> > Deepak
> >>
>
+
Patrick Hunt 2012-05-17, 23:29