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

Switch to Threaded View
HBase, mail # user - how to reduce zookeeper.out


Copy link to this message
-
Re: 答复: how to reduce zookeeper.out
Jean-Marc Spaggiari 2013-03-30, 14:57
This is the kind of info I have on my ZK.out:
2013-03-30 10:54:33,605 [myid:1] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@213] -
Accepted socket connection from /192.168.23.10:40812
2013-03-30 10:54:33,685 [myid:1] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@838] -
Client attempting to establish new session at /192.168.23.10:40812
2013-03-30 10:54:33,706 [myid:1] - INFO
[CommitProcessor:1:ZooKeeperServer@604] - Established session
0x13d783198e46fcf with negotiated timeout 40000 for client
/192.168.23.10:40812
2013-03-30 10:54:35,148 [myid:1] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@213] -
Accepted socket connection from /192.168.23.5:34998
2013-03-30 10:54:35,160 [myid:1] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:NIOServerCnxnFactory@213] -
Accepted socket connection from /192.168.23.5:34999
2013-03-30 10:54:35,177 [myid:1] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@838] -
Client attempting to establish new session at /192.168.23.5:34998
2013-03-30 10:54:35,178 [myid:1] - INFO
[NIOServerCxn.Factory:0.0.0.0/0.0.0.0:2181:ZooKeeperServer@838] -
Client attempting to establish new session at /192.168.23.5:34999
2013-03-30 10:54:35,195 [myid:1] - INFO
[CommitProcessor:1:ZooKeeperServer@604] - Established session
0x13d783198e46fd0 with negotiated timeout 40000 for client
/192.168.23.5:34998
2013-03-30 10:54:35,207 [myid:1] - INFO
[CommitProcessor:1:ZooKeeperServer@604] - Established session
0x13d783198e46fd1 with negotiated timeout 40000 for client
/192.168.23.5:34999
also, I just tried to rm and touch and it's now writing on the file
anymore, so it's not a good solution.

I just stopped and restarted it and it's now writing on the file
again. I did that because I have 3 ZK servers... So it should be safe.

I have a MR job running right now, and it does seems it has impacted
it. But I have not check all the logs yet.

JM

2013/3/30 huaxiang <[EMAIL PROTECTED]>:
> KEVIN,
>    This file is usualy could be found under "bin" directory, that is
> "bin/zookeeper.out".
>
>
>
>
> -----邮件原件-----
> 发件人: Kevin O'dell [mailto:[EMAIL PROTECTED]]
> 发送时间: 2013年3月30日 22:47
> 收件人: [EMAIL PROTECTED]
> 主题: Re: how to reduce zookeeper.out
>
> Hua,
>
>   That is no fun.  Another question though, I just checked my clusters and I
> don't even have a zookeeper.out.  What messages do you have filing this file
> up?
>
> On Sat, Mar 30, 2013 at 10:42 AM, hua beatls <[EMAIL PROTECTED]> wrote:
>
>> Hi Kevin,
>>      seems there is no rolling mechanizm for the zookeeper.out.
>>
>>
>>
>>
>> On Sat, Mar 30, 2013 at 10:31 PM, Kevin O'dell
>> <[EMAIL PROTECTED]
>> >wrote:
>>
>> > Hi Hua,
>> >
>> >   I believe(don't quote me) that you can use the rolling file
>> > appender to set the files to a max size.  I know HBase does this,
>> > but I am not sure about ZK.
>> >
>> > On Sat, Mar 30, 2013 at 10:22 AM, Jean-Marc Spaggiari <
>> > [EMAIL PROTECTED]> wrote:
>> >
>> > > Hi,
>> > >
>> > > If you are in production you should have more than 1 ZK server, right?
>> > > Like, 3 or more?
>> > >
>> > > Then you can just shut one down, the others mike elect a new leader.
>> > > Then you clear the zookeeper.out file, and start it back?
>> > >
>> > > JM
>> > >
>> > > 2013/3/30 hua beatls <[EMAIL PROTECTED]>:
>> > > > Hi,
>> > > >     zookeeper.out is too large, how to reduce this file? it is a
>> online
>> > > > system.
>> > > >
>> > > >    Best R.
>> > > >
>> > > >   beatls
>> > >
>> >
>> >
>> >
>> > --
>> > Kevin O'Dell
>> > Customer Operations Engineer, Cloudera
>> >
>>
>
>
>
> --
> Kevin O'Dell
> Customer Operations Engineer, Cloudera
>