Return-Path: X-Original-To: apmail-hbase-user-archive@www.apache.org Delivered-To: apmail-hbase-user-archive@www.apache.org Received: from mail.apache.org (hermes.apache.org [140.211.11.3]) by minotaur.apache.org (Postfix) with SMTP id 01CE9D301 for ; Thu, 29 Nov 2012 01:53:47 +0000 (UTC) Received: (qmail 91600 invoked by uid 500); 29 Nov 2012 01:53:44 -0000 Delivered-To: apmail-hbase-user-archive@hbase.apache.org Received: (qmail 91540 invoked by uid 500); 29 Nov 2012 01:53:44 -0000 Mailing-List: contact user-help@hbase.apache.org; run by ezmlm Precedence: bulk List-Help: List-Unsubscribe: List-Post: List-Id: Reply-To: user@hbase.apache.org Delivered-To: mailing list user@hbase.apache.org Received: (qmail 91532 invoked by uid 99); 29 Nov 2012 01:53:44 -0000 Received: from athena.apache.org (HELO athena.apache.org) (140.211.11.136) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 29 Nov 2012 01:53:44 +0000 X-ASF-Spam-Status: No, hits=-0.0 required=5.0 tests=RCVD_IN_DNSWL_NONE,SPF_PASS X-Spam-Check-By: apache.org Received-SPF: pass (athena.apache.org: domain of michael_segel@hotmail.com designates 65.55.111.100 as permitted sender) Received: from [65.55.111.100] (HELO blu0-omc2-s25.blu0.hotmail.com) (65.55.111.100) by apache.org (qpsmtpd/0.29) with ESMTP; Thu, 29 Nov 2012 01:53:39 +0000 Received: from BLU0-SMTP308 ([65.55.111.73]) by blu0-omc2-s25.blu0.hotmail.com with Microsoft SMTPSVC(6.0.3790.4675); Wed, 28 Nov 2012 17:53:17 -0800 X-Originating-IP: [173.15.87.37] X-EIP: [/U+j71HJ7YkDMArkDxeth11+la+mHb4L] X-Originating-Email: [michael_segel@hotmail.com] Message-ID: Received: from [192.168.0.104] ([173.15.87.37]) by BLU0-SMTP308.blu0.hotmail.com over TLS secured channel with Microsoft SMTPSVC(6.0.3790.4675); Wed, 28 Nov 2012 17:53:16 -0800 Content-Type: text/plain; charset="iso-8859-1" MIME-Version: 1.0 (Mac OS X Mail 6.2 \(1499\)) Subject: Re: recommended nodes From: Michael Segel In-Reply-To: Date: Wed, 28 Nov 2012 19:53:14 -0600 Content-Transfer-Encoding: quoted-printable References: <50B37737.7050409@uci.cu> To: user@hbase.apache.org X-Mailer: Apple Mail (2.1499) X-OriginalArrivalTime: 29 Nov 2012 01:53:16.0842 (UTC) FILETIME=[4CC608A0:01CDCDD4] X-Virus-Checked: Checked by ClamAV on apache.org Ok, just a caveat. I am discussing MapR as part of a complete response. As Mohit posted = MapR takes the raw device for their MapR File System.=20 They do stripe on their own within what they call a volume.=20 But going back to Apache...=20 You can stripe drives, however I wouldn't recommend it. I don't think = the performance gains would really matter.=20 You're going to end up getting blocked first by disk i/o, then your = controller card, then your network... assuming 10GBe.=20 With only 2 disks on an 8 core system, you will hit disk i/o first and = then you'll watch your CPU Wait I/O climb.=20 HTH -Mike On Nov 28, 2012, at 7:28 PM, Jean-Marc Spaggiari = wrote: > Hi Mike, >=20 > Why not using LVM with MapR? Since LVM is reading from 2 drives almost > at the same time, it should be better than RAID0 or a single drive, > no? >=20 > 2012/11/28, Michael Segel : >> Just a couple of things. >>=20 >> I'm neutral on the use of LVMs. Some would point out that there's = some >> overhead, but on the flip side, it can make managing the machines = easier. >> If you're using MapR, you don't want to use LVMs but raw devices. >>=20 >> In terms of GC, its going to depend on the heap size and not the = total >> memory. With respect to HBase. ... MSLABS is the way to go. >>=20 >>=20 >> On Nov 28, 2012, at 12:05 PM, Jean-Marc Spaggiari = >> wrote: >>=20 >>> Hi Gregory, >>>=20 >>> I founs this about LVM: >>> -> http://blog.andrew.net.au/2006/08/09 >>> -> = http://www.phoronix.com/scan.php?page=3Darticle&item=3Dfedora_15_lvm&num=3D= 2 >>>=20 >>> Seems that performances are still correct with it. I will most >>> probably give it a try and bench that too... I have one new hard = drive >>> which should arrived tomorrow. Perfect timing ;) >>>=20 >>>=20 >>>=20 >>> JM >>>=20 >>> 2012/11/28, Mohit Anchlia : >>>>=20 >>>>=20 >>>>=20 >>>>=20 >>>> On Nov 28, 2012, at 9:07 AM, Adrien Mogenet = >>>> wrote: >>>>=20 >>>>> Does HBase really benefit from 64 GB of RAM since allocating too = large >>>>> heap >>>>> might increase GC time ? >>>>>=20 >>>> Benefit you get is from OS cache >>>>> Another question : why not RAID 0, in order to aggregate disk = bandwidth >>>>> ? >>>>> (and thus keep 3x replication factor) >>>>>=20 >>>>>=20 >>>>> On Wed, Nov 28, 2012 at 5:58 PM, Michael Segel >>>>> wrote: >>>>>=20 >>>>>> Sorry, >>>>>>=20 >>>>>> I need to clarify. >>>>>>=20 >>>>>> 4GB per physical core is a good starting point. >>>>>> So with 2 quad core chips, that is going to be 32GB. >>>>>>=20 >>>>>> IMHO that's a minimum. If you go with HBase, you will want more. >>>>>> (Actually >>>>>> you will need more.) The next logical jump would be to 48 or = 64GB. >>>>>>=20 >>>>>> If we start to price out memory, depending on vendor, your = company's >>>>>> procurement, there really isn't much of a price difference in = terms >>>>>> of >>>>>> 32,48, or 64 GB. >>>>>> Note that it also depends on the chips themselves. Also you need = to >>>>>> see >>>>>> how many memory channels exist in the mother board. You may need = to >>>>>> buy >>>>>> in >>>>>> pairs or triplets. Your hardware vendor can help you. (Also you = need >>>>>> to >>>>>> keep an eye on your hardware vendor. Sometimes they will give you >>>>>> higher >>>>>> density chips that are going to be more expensive...) ;-) >>>>>>=20 >>>>>> I tend to like having extra memory from the start. >>>>>> It gives you a bit more freedom and also protects you from 'fat' = code. >>>>>>=20 >>>>>> Looking at YARN... you will need more memory too. >>>>>>=20 >>>>>>=20 >>>>>> With respect to the hard drives... >>>>>>=20 >>>>>> The best recommendation is to keep the drives as JBOD and then = use 3x >>>>>> replication. >>>>>> In this case, make sure that the disk controller cards can handle >>>>>> JBOD. >>>>>> (Some don't support JBOD out of the box) >>>>>>=20 >>>>>> With respect to RAID... >>>>>>=20 >>>>>> If you are running MapR, no need for RAID. >>>>>> If you are running an Apache derivative, you could use RAID 1. = Then >>>>>> cut >>>>>> your replication to 2X. This makes it easier to manage drive = failures. >>>>>> (Its not the norm, but it works...) In some clusters, they are = using >>>>>> appliances like Net App's e series where the machines see the = drives >>>>>> as >>>>>> local attached storage and I think the appliances themselves are = using >>>>>> RAID. I haven't played with this configuration, however it could = make >>>>>> sense and its a valid design. >>>>>>=20 >>>>>> HTH >>>>>>=20 >>>>>> -Mike >>>>>>=20 >>>>>> On Nov 28, 2012, at 10:33 AM, Jean-Marc Spaggiari >>>>>> >>>>>> wrote: >>>>>>=20 >>>>>>> Hi Mike, >>>>>>>=20 >>>>>>> Thanks for all those details! >>>>>>>=20 >>>>>>> So to simplify the equation, for 16 virtual cores we need 48 to = 64GB. >>>>>>> Which mean 3 to 4GB per core. So with quad cores, 12GB to 16GB = are a >>>>>>> good start? Or I simplified it to much? >>>>>>>=20 >>>>>>> Regarding the hard drives. If you add more than one drive, do = you >>>>>>> need >>>>>>> to build them on RAID or similar systems? Or can Hadoop/HBase be >>>>>>> configured to use more than one drive? >>>>>>>=20 >>>>>>> Thanks, >>>>>>>=20 >>>>>>> JM >>>>>>>=20 >>>>>>> 2012/11/27, Michael Segel : >>>>>>>>=20 >>>>>>>> OK... I don't know why Cloudera is so hung up on 32GB. ;-) [Its = an >>>>>> inside >>>>>>>> joke ...] >>>>>>>>=20 >>>>>>>> So here's the problem... >>>>>>>>=20 >>>>>>>> By default, your child processes in a map/reduce job get a = default >>>>>> 512MB. >>>>>>>> The majority of the time, this gets raised to 1GB. >>>>>>>>=20 >>>>>>>> 8 cores (dual quad cores) shows up at 16 virtual processors in >>>>>>>> Linux. >>>>>> (Note: >>>>>>>> This is why when people talk about the number of cores, you = have to >>>>>> specify >>>>>>>> physical cores or logical cores....) >>>>>>>>=20 >>>>>>>> So if you were to over subscribe and have lets say 12 mappers = and >>>>>>>> 12 >>>>>>>> reducers, that's 24 slots. Which means that you would need 24GB = of >>>>>> memory >>>>>>>> reserved just for the child processes. This would leave 8GB for = DN, >>>>>>>> TT >>>>>> and >>>>>>>> the rest of the linux OS processes. >>>>>>>>=20 >>>>>>>> Can you live with that? Sure. >>>>>>>> Now add in R, HBase, Impala, or some other set of tools on top = of >>>>>>>> the >>>>>>>> cluster. >>>>>>>>=20 >>>>>>>> Ooops! Now you are in trouble because you will swap. >>>>>>>> Also adding in R, you may want to bump up those child procs = from 1GB >>>>>>>> to >>>>>> 2 >>>>>>>> GB. That means the 24 slots would now require 48GB. Now you = have >>>>>>>> swap >>>>>> and >>>>>>>> if that happens you will see HBase in a cascading failure. >>>>>>>>=20 >>>>>>>> So while you can do a rolling restart with the changed = configuration >>>>>>>> (reducing the number of mappers and reducers) you end up with = less >>>>>>>> slots >>>>>>>> which will mean in longer run time for your jobs. (Less slots = =3D=3D >>>>>>>> less >>>>>>>> parallelism ) >>>>>>>>=20 >>>>>>>> Looking at the price of memory... you can get 48GB or even 64GB = for >>>>>> around >>>>>>>> the same price point. (8GB chips) >>>>>>>>=20 >>>>>>>> And I didn't even talk about adding SOLR either again a memory >>>>>>>> hog... >>>>>> ;-) >>>>>>>>=20 >>>>>>>> Note that I matched the number of mappers w reducers. You could = go >>>>>>>> with >>>>>>>> fewer reducers if you want. I tend to recommend a ratio of 2:1 >>>>>>>> mappers >>>>>> to >>>>>>>> reducers, depending on the work flow.... >>>>>>>>=20 >>>>>>>> As to the disks... no 7200 SATA III drives are fine. SATA III >>>>>>>> interface >>>>>> is >>>>>>>> pretty much available in the new kit being shipped. >>>>>>>> Its just that you don't have enough drives. 8 cores should be 8 >>>>>> spindles if >>>>>>>> available. >>>>>>>> Otherwise you end up seeing your CPU load climb on wait states = as >>>>>>>> the >>>>>>>> processes wait for the disk i/o to catch up. >>>>>>>>=20 >>>>>>>> I mean you could build out a cluster w 4 x 3 3.5" 2TB drives in = a 1 >>>>>>>> U >>>>>>>> chassis based on price. You're making a trade off and you = should be >>>>>> aware of >>>>>>>> the performance hit you will take. >>>>>>>>=20 >>>>>>>> HTH >>>>>>>>=20 >>>>>>>> -Mike >>>>>>>>=20 >>>>>>>> On Nov 27, 2012, at 1:52 PM, Jean-Marc Spaggiari < >>>>>> jean-marc@spaggiari.org> >>>>>>>> wrote: >>>>>>>>=20 >>>>>>>>> Hi Michael, >>>>>>>>>=20 >>>>>>>>> so are you recommanding 32Gb per node? >>>>>>>>>=20 >>>>>>>>> What about the disks? SATA drives are to slow? >>>>>>>>>=20 >>>>>>>>> JM >>>>>>>>>=20 >>>>>>>>> 2012/11/26, Michael Segel : >>>>>>>>>> Uhm, those specs are actually now out of date. >>>>>>>>>>=20 >>>>>>>>>> If you're running HBase, or want to also run R on top of = Hadoop, >>>>>>>>>> you >>>>>>>>>> will >>>>>>>>>> need to add more memory. >>>>>>>>>> Also forget 1GBe got 10GBe, and w 2 SATA drives, you will be = disk >>>>>>>>>> i/o >>>>>>>>>> bound >>>>>>>>>> way too quickly. >>>>>>>>>>=20 >>>>>>>>>>=20 >>>>>>>>>> On Nov 26, 2012, at 8:05 AM, Marcos Ortiz = wrote: >>>>>>>>>>=20 >>>>>>>>>>> Are you asking about hardware recommendations? >>>>>>>>>>> Eric Sammer on his "Hadoop Operations" book, did a great job >>>>>>>>>>> about >>>>>>>>>>> this: >>>>>>>>>>> For middle size clusters (until 300 nodes): >>>>>>>>>>> Processor: A dual quad-core 2.6 Ghz >>>>>>>>>>> RAM: 24 GB DDR3 >>>>>>>>>>> Dual 1 Gb Ethernet NICs >>>>>>>>>>> a SAS drive controller >>>>>>>>>>> at least two SATA II drives in a JBOD configuration >>>>>>>>>>>=20 >>>>>>>>>>> The replication factor depends heavily of the primary use of = your >>>>>>>>>>> cluster. >>>>>>>>>>>=20 >>>>>>>>>>> On 11/26/2012 08:53 AM, David Charle wrote: >>>>>>>>>>>> hi >>>>>>>>>>>>=20 >>>>>>>>>>>> what's the recommended nodes for NN, hmaster and zk nodes = for a >>>>>> larger >>>>>>>>>>>> cluster, lets say 50-100+ >>>>>>>>>>>>=20 >>>>>>>>>>>> also, what would be the ideal replication factor for larger >>>>>>>>>>>> clusters >>>>>>>>>>>> when >>>>>>>>>>>> u have 3-4 racks ? >>>>>>>>>>>>=20 >>>>>>>>>>>> -- >>>>>>>>>>>> David >>>>>>>>>>>> 10mo. ANIVERSARIO DE LA CREACION DE LA UNIVERSIDAD DE LAS >>>>>>>>>>>> CIENCIAS >>>>>>>>>>>> INFORMATICAS... >>>>>>>>>>>> CONECTADOS AL FUTURO, CONECTADOS A LA REVOLUCION >>>>>>>>>>>>=20 >>>>>>>>>>>> http://www.uci.cu >>>>>>>>>>>> http://www.facebook.com/universidad.uci >>>>>>>>>>>> http://www.flickr.com/photos/universidad_uci >>>>>>>>>>>=20 >>>>>>>>>>> -- >>>>>>>>>>>=20 >>>>>>>>>>> Marcos Luis Ort=EDz Valmaseda >>>>>>>>>>> about.me/marcosortiz >>>>>>>>>>> @marcosluis2186 >>>>>>>>>>>=20 >>>>>>>>>>>=20 >>>>>>>>>>>=20 >>>>>>>>>>> 10mo. ANIVERSARIO DE LA CREACION DE LA UNIVERSIDAD DE LAS >>>>>>>>>>> CIENCIAS >>>>>>>>>>> INFORMATICAS... >>>>>>>>>>> CONECTADOS AL FUTURO, CONECTADOS A LA REVOLUCION >>>>>>>>>>>=20 >>>>>>>>>>> http://www.uci.cu >>>>>>>>>>> http://www.facebook.com/universidad.uci >>>>>>>>>>> http://www.flickr.com/photos/universidad_uci >>>>>=20 >>>>>=20 >>>>> -- >>>>> Adrien Mogenet >>>>> 06.59.16.64.22 >>>>> http://www.mogenet.me >>>>=20 >>>=20 >>=20 >>=20 >=20