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 Plain View
MapReduce >> mail # user >> Re: What should I do with a 48-node cluster


+
Mark Kerzner 2012-12-21, 02:07
+
Lance Norskog 2012-12-21, 19:37
+
Mark Kerzner 2012-12-21, 20:03
+
Edward Capriolo 2012-12-21, 23:50
+
Mark Kerzner 2012-12-23, 00:53
Copy link to this message
-
Re: What should I do with a 48-node cluster
You do not absolutely need more ram.  You do not know your workload yet. A
standard hadoop machine has 8 disks 16 GB RAM, 8 cores.

In the old days, you would dedicate map slots and reduce slots map 3 map 1
reduce in your case. Give each of them 256 RAM for child jvm ops. So you
needed more ram in that case that you have 8 cores, but you do not.

In the end blades are not the ideal hadoop machine because users usually
want many disks for lots of IO, but it is ok for kicking around.

On Sat, Dec 22, 2012 at 7:53 PM, Mark Kerzner <[EMAIL PROTECTED]>wrote:

> Edward, thank you for the practical recommendations. I am going to visit
> the cluster in its current home in a few days, and I will keep this in
> mind. Meanwhile, my specs are below
>
> 48 HP 1U blades, each has two 2.44 GHz. Dual core AMD Opterons with Cisco
> Infiniband NICs, 4GB RAM
>
> 1 HP cluster controller with SCSI controller
>
> 1 HP RSA20 storage array with approx 1Tb of storage
>
> Cisco Infiniband 20Gbit optical network router
>
> In Compaq racks with four 30 amp 220 volt circuits
>
> All wiring and cabling.
>
> I am worried about 4 GB RAM on data nodes not being enough. Upgrading the
> master nodes is bearable, but any memory upgrade on the complete cluster
> will sure cost, when multiplied by 50.
>
> Thank you. Sincerely,
> Mark
>
> On Fri, Dec 21, 2012 at 5:50 PM, Edward Capriolo <[EMAIL PROTECTED]>wrote:
>
>> Three year old blade center is ok. A three year old blade is probably a
>> 64 bit machine. 2 to 4 gb RAM 2 SCSI disks. Maybe two socket two core. Two
>> blade centers is about 8u or a quarter cabinet and you can find a hosting
>> provider in your price range.
>>
>> Especially if you can get the hardware at a low initial cost you crush
>> the cloud providers. Buying your own gear takes about a year to recoup
>> costs over amazons pay per use model.
>>
>> Blade centers are usually 20 to 30 amp fully loaded though so if your
>> crushing word count at home your power bill is gonna get $.
>>
>>
>>
>> On Friday, December 21, 2012, Mark Kerzner <[EMAIL PROTECTED]>
>> wrote:
>> > True!
>> >
>> > I am thinking of either my (small) office, or actually hosting for
>> under $500/month.
>> >
>> > On Fri, Dec 21, 2012 at 1:37 PM, Lance Norskog <[EMAIL PROTECTED]>
>> wrote:
>> >>
>> >> You will also be raided by the DEA- too much power for a residence.
>> >>
>> >> On 12/20/2012 07:56 AM, Ted Dunning wrote:
>> >>
>> >>
>> >>
>> >> On Thu, Dec 20, 2012 at 7:38 AM, Michael Segel <
>> [EMAIL PROTECTED]> wrote:
>> >>>
>> >>> While Ted ignores that the world is going to end before X-Mas, he
>> does hit the crux of the matter head on.
>> >>> If you don't have a place to put it, the cost of setting it up would
>> kill you, not to mention that you can get newer hardware which is better
>> suited for less.
>> >>> Having said that... if you live in the frozen tundra like Montana, or
>> some place like ... er Canada or Siberia... , it may make more sense to use
>> it to heat your home with it.
>> >>> Just think of the side benefits from all that potential additional
>> compute power....  :-P
>> >>
>> >> I can say from experience that the sound of a bunch of servers in a
>> home setting is a novel one that is probably unlike anything you have known
>> before.
>> >> If you haven't experienced that, then taking on these servers could be
>> classified as novelty seeking behavior.
>> >
>> >
>>
>
>
+
Mark Kerzner 2012-12-23, 01:43
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