hadoop-zookeeper-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Mahadev Konar <maha...@yahoo-inc.com>
Subject Re: Server exception when closing session
Date Sat, 23 Jan 2010 01:48:50 GMT
HI Josh,
  The server latency does seem huge. What os and hardware are you running it
on? What is usage model of zookeeper? How much memory are you allocating to
the server? 
The debug well exacerbate the problem.
A dedicated disk means the following:
Zookeeper has snapshots and transaction logs. The datadir is the directory
that stores the transaction logs. Its highly recommended that this directory
be on a separate disk that isnt being used by any other process. The
snapshots can sit on a disk that is being used by the OS and can be shared.

Also, Pat ran some tests for serve lantecies at:

http://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview

You can take a look at that as well and see what the expected performance
should be for your workload.

Thanks
mahadev


On 1/22/10 5:40 PM, "Josh Scheid" <jscheid@velocetechnologies.com> wrote:

> On Fri, Jan 22, 2010 at 17:22, Mahadev Konar <mahadev@yahoo-inc.com> wrote:
>>  This warning is not of any concern.
> 
> OK.  I'm used to warnings being things that must be addressed.  I'll
> ignore this one in the future.
> 
>> Just a quick question, is there any reason for you to runn the server on a
>> DEBUG level?
> 
> We're having issues with server latency.  Client default timeout of
> 10000ms gets hit.  I saw a "stat" output showing a 16s max latency
> today.
> Is DEBUG going to exacerbate that?  Of the recommendations I've seen,
> the one I can't yet follow is a dedicated disk: dataDir is in the root
> partition of the server right now.
> 
> -Josh


Mime
View raw message