Home | About | Sematext search-lucene.com search-hadoop.com
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB
 Search Hadoop and all its subprojects:

Switch to Threaded View
MapReduce >> mail # user >> Can I move block data directly?


Copy link to this message
-
Re: Can I move block data directly?
bq. I'd also ensure the ownership of the block files are intact.
Hi Harsh,

what's mean "ensure the ownership of the block files are intact"?

and I want to ask more.

for my understand, after I restart the data node daemon, block report
should tell NN all blocks owned by this DN, and block scnaner can remember
all blocks' structure on local, so the block file owner ship would be
confirmed at the starting period

and even if some pieces of blk_ files losts, then NN an find it's under
replicated, am I right?  Thanks.
On Mon, Jul 8, 2013 at 2:07 PM, Azuryy Yu <[EMAIL PROTECTED]> wrote:

> Thanks Harsh, always detailed answers each time.
>
> Yes, this is an unsupported scenarios, Load balancer is very slow even
> after I set bandwidthPerSec  to a large value, so I want to take this way
> to slove the problem quickly.
>
>
>
>
>
>
> On Mon, Jul 8, 2013 at 1:46 PM, Viral Bajaria <[EMAIL PROTECTED]>wrote:
>
>> Out of curiosity. Besides the bandwidthPerSec and threshold what other
>> parameters are tuneable ?
>>
>> Thanks,
>> Viral
>>
>> On Sun, Jul 7, 2013 at 10:39 PM, Harsh J <[EMAIL PROTECTED]> wrote:
>>
>>> If balancer isn't cutting it out for you with stock defaults, you
>>> should consider tuning that than doing these unsupported scenarios.
>>>
>>
>>
>>
>
NEW: Monitor These Apps!
elasticsearch, apache solr, apache hbase, hadoop, redis, casssandra, amazon cloudwatch, mysql, memcached, apache kafka, apache zookeeper, apache storm, ubuntu, centOS, red hat, debian, puppet labs, java, senseiDB