hadoop-hdfs-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Joey Echeverria <j...@cloudera.com>
Subject Re: Hadoop Namenode Failover
Date Mon, 02 Apr 2012 14:13:17 GMT
Automatic failover is coming in a future release of the 0.23.x release
series. I'd recommend waiting until a release with it built in as
trying to build automatic failover with proper fencing is not easy.

What's your use case? If you're using HDFS so that you can do batch
processing with MapReduce, minutes of down time shouldn't be a deal
breaker. Also, in practice NN failures are rare.

-Joey

On Thu, Mar 22, 2012 at 3:11 PM, Tibor Korocz <tkorocz@gmail.com> wrote:
> Hi Bejoy,
>
> Thank you for your answer.
> - I read about Secondary Name Node, but it is not realtime solution. In this
> solution are some minutes delay...
> - With NFS how can i make it, the Data Node automatic change to another Name
> Node..?
>   Maybe this: http://wiki.apache.org/hadoop/NameNodeFailover or i found
> this https://issues.apache.org/jira/browse/HDFS-976
>
> Somebody has experience with Avatar Node?
>
> Tibi
>
>
> 2012/3/22 Bejoy Ks <bejoy.hadoop@gmail.com>
>>
>> Hi Tibi
>>       To recover from a Name Node failure, the following are
>> the suggested approaches as of now
>>
>> - Maintain a SNN which does periodic check pointing
>>
>> http://hadoop.apache.org/common/docs/current/hdfs_user_guide.html#Secondary+NameNode
>> - Make sure you configure a remote NFS mount, so you have a remote copy of
>> fs image on another server even if you fully lose out the running NN
>>
>> There is some awesome work going on for HA within hadoop project itself,
>> for details
>> https://issues.apache.org/jira/browse/HDFS-1623
>>
>> Regards
>> Bejoy KS
>>
>>
>> On Thu, Mar 22, 2012 at 7:56 PM, Tibor Korocz <tkorocz@gmail.com> wrote:
>>>
>>> Hi,
>>>
>>> I am new in Hadoop. I created a four nodes cluster 3 slaves, and 1 master
>>> (this is onyl the test system).
>>> It works fine, i used this howto:
>>> http://www.michael-noll.com/tutorials/running-hadoop-on-ubuntu-linux-multi-node-cluster/
>>> .
>>> My question is, what is the best solution to make the master(namenode)
>>> fail over, i read a lot, but i dont now what is the best.
>>> I found this howto:
>>> http://www.cloudera.com/blog/2009/07/hadoop-ha-configuration/ , but if it is
>>> possible, i do not want to use DRBD.
>>> I hope somebody can help me. Sorry my english.... :)
>>>
>>> Thanks.
>>> Tibi
>>
>>
>



-- 
Joey Echeverria
Solutions Architect
Cloudera, Inc.

Mime
View raw message