unsubscribe

[Levin Ding]

Pls unsubscribe me. Thx.
在 2013-3-16，上午3:03，kishore raju  写道：

> HI,
> 
>  We are having an issue where multiple Task Trackers are running out of 
> memory. I have collected HeapDump on those TaskTrackers to analyze further. 
> They are currently running with 1GB Heap. we are planning to bump it to 2GB, 
> Is there a way that we can find  which Job is causing this OOM on TT's ?
> 
> Any help is appreciated.
> 
> 
> -Thanks
>  kishore 

Re: unsubscribe

[Levin ding]

unsubscribe

Re: HDFS with JBOD configuration

[Levin ding]

unsubscribe


2014-04-03 20:54 GMT+08:00 anuj maurice :

> Use dfs.datanode.failed.volumes.tolerated
>
> http://hadoop.apache.org/docs/r2.3.0/hadoop-project-dist/hadoop-hdfs/hdfs-default.xml
>
>
> On Thu, Apr 3, 2014 at 6:07 PM, Stanley Shi  wrote:
>
>> AFAIK, data node will be dead since it cannot handle one disk failure (
>> if it cannot write data to a disk, it will fail).
>> But not sure if this situation has changed.
>>
>> Regards,
>> *Stanley Shi,*
>>
>>
>>
>> On Thu, Apr 3, 2014 at 7:29 PM, Bharath Kumar wrote:
>>
>>>
>>> Hi All,
>>>  I have a query with respect to JBOD configuration . Suppose in
>>> my data node if one of the
>>> disk crashes in JBOD configuration will the entire data node be shown as
>>> dead/unavailable ?
>>> --
>>> Warm Regards,
>>>  *Bharath Kumar *
>>>
>>>
>>>
>>
>>
>
>
> --
> regards ,
> Anuj Maurice
>

unsubscribe

[Levin ding]

pls unsubscribe me, thanks.

unsubscribe

[Levin ding]

unsubscribe