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
HBase >> mail # user >> HBASE -- Regionserver and QuorumPeer ?


Copy link to this message
-
Re: HBASE -- Regionserver and QuorumPeer ?
Well...

I wasn't sure if St.Ack was displeased by my comments on Andrew's response, or my references to KISS where the second S is stupid, reference to 'dead hookers' or reference to drugs.
I was just covering my bases. :-)

With respect to Andrew's response, I saw something that I wasn't sure if I was reading too much in to his response.  Hence my start with that I may be losing it because I was probably reading something in to his response that he may not have intended.
I guess its a problem many of us have, myself included, where we are sometimes intentionally vague in our response.

There are times when someone asks a question, the response is that they shouldn't do X, that while its not a good idea to do something, its still theoretically possible to do.
In this case running a RS and ZK on the same node.  Yes, it could be done with the proper configuration where you isolate your disk I/O as much as possible between ZK and the RS. However the better solution is to run the ZK along with the JT, NN, HM and even SN on the same node. (For a small dev cluster.)

Another case in point is that we see things taken out of context. As an example, there was a presentation by Facebook I think... where they run their HBase on nodes where they don't run TT. In context, this could make a lot of sense when they are using HBase to deliver real time response to an app outside of the cluster, and are not using it as part of a M/R job. The problem is that someone sees this and takes it out of context saying that FB does it and the best way to run HBase is to not run it on the same nodes you have TT running.  (Data Locality? Forget about it...)

Note I don't believe that this is what the FB presentation was suggesting except in their specific solution.

In another thread ,  someone was asking for help because they were having problems with their cluster. One node was in India, Two were in the US. The response was along the idea that its not a good thing to do this.   While I agree with the response, I have to wonder if it shouldn't have been worded more strongly. We aren't saying it can't be done, we're saying that its not something we'd recommend. I don't know if that's a strong enough response to really discourage an OP from actually doing it.

Is that a better explanation?

On Jul 2, 2012, at 6:53 AM, Mohammad Tariq wrote:

> What kind of explanation is this???????????
>
> Regards,
>    Mohammad Tariq
>
>
> On Mon, Jul 2, 2012 at 5:10 PM, Michael Segel <[EMAIL PROTECTED]> wrote:
>> Sorry St. Ack,
>>
>> Which is why I said that I was losing it...
>>
>> The entire quote was...
>> "On Sun, Jul 1, 2012 at 2:05 PM, Jay Wilson
>> <[EMAIL PROTECTED]> wrote:
>>> Can a regionserver and quorumpeer reside on the same node?
>>
>> It can, but you want to consider how disk is allocated in the cluster.
>>
>> A typical and recommended configuration is HBase RegionServer and HDFS
>> DataNode colocated on the nodes. The DataNode will use locally
>> attached disk to store and serve blocks.
>> "
>>
>> Looking at and parsing this you have two things...
>>
>> 1) When reading the 'A typical and recommended configuration...' can imply that its possible while not recommended to try and run an HBase RS while not running a DN service on the same node.
>>
>> 2) "It can, but you want to consider how disk is allocated in the cluster."
>> While on a single machine running as a pseudo cluster is one thing, running a fully distributed cluster is another.
>>
>>
>> I am not finding fault with what Andy was saying. The problem is that we tend not to use stronger language when discussing these topics.  And my point wasn't just on this topic but others posts where we say 'not a good idea' yet someone still pursues the idea until there's a chorus of saying not to do something.  I'm not faulting the poster because he wasn't and isn't the only one who does this... We see it all the time where someone goes down the wrong path, and is looking for a quick solution, rather than following the recommendation.
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