Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
HBase, mail # user - Acceptable CPU_WIO % ?


+
Jean-Marc Spaggiari 2013-02-08, 01:19
+
Kevin Odell 2013-02-08, 01:43
+
Jean-Marc Spaggiari 2013-02-08, 02:00
+
Kevin Odell 2013-02-08, 02:21
+
Jean-Marc Spaggiari 2013-02-08, 02:50
+
Kevin Odell 2013-02-08, 02:57
+
Jean-Marc Spaggiari 2013-02-08, 03:15
+
Azuryy Yu 2013-02-08, 03:23
+
Kevin Odell 2013-02-08, 13:56
+
Jean-Marc Spaggiari 2013-02-08, 15:43
+
Kevin Odell 2013-02-08, 16:37
Copy link to this message
-
Re: Acceptable CPU_WIO % ?
Jean-Marc Spaggiari 2013-02-09, 16:13
Perfect, thanks Kevin.

Looking at SAR this morning, I can see that I'm sometime reaching
300tps, and spiked at 80% WIO... That will cost me 5 new additional
hard drives :(

I’m not sure I will have time to install them all today, but as soon
as it’s done I will give you some news.

JM
2013/2/8, Kevin O'dell <[EMAIL PROTECTED]>:
> JM,
>
>   Basically, you will have to replace failed disk and rebuild RAID0 since
> the other half of the data is worthless.  There is not a real recommended
> value, but anything under 150 - 200 would make me more comfortable.
>
> On Fri, Feb 8, 2013 at 10:43 AM, Jean-Marc Spaggiari <
> [EMAIL PROTECTED]> wrote:
>
>> Hi Kevin,
>>
>> I think it will take time before I get a chance to have 5 drives in
>> the same server, so I will see at that time to test RAID5.
>>
>> I'm going to add one drive per server today or tomorrow to try to
>> improve that. What IOPs should I try to have? 100? Less? It will all
>> be SATA3 drives and I will configure all in RAID0.
>>
>> It doesn't seems to me to be an issue to lose one node, since data
>> will be replicated everywhere else. I will "simply" have to replace
>> the failing disk and restart the node, no?
>>
>> JM
>>
>> 2013/2/8, Kevin O'dell <[EMAIL PROTECTED]>:
>> > Azuryy,
>> >
>> >   The main reason to recommend against RAID is that it is slow and it
>> adds
>> > redundancy that we already have in Hadoop.  RAID0 is another story as
>> long
>> > as all of the drives are healthy and you don't mind losing the whole
>> volume
>> > if you lose one drive.
>> >
>> > JM,
>> >
>> >   I would not even waste my time testing RAID5 or RAID6(unless it is
>> > just
>> > for educational purposes :) ).  200+ IOPs consistently on one SATA
>> > drive
>> is
>> > pretty high, that would explain your high I/O wait time.  If your use
>> case
>> > allows for you to lose the whole node, there is not a good reason for
>> > you
>> > to shy away from RAID0.  Please let us know how this plays out with
>> > your
>> > environment.
>> >
>> > On Thu, Feb 7, 2013 at 10:23 PM, Azuryy Yu <[EMAIL PROTECTED]> wrote:
>> >
>> >> JM,
>> >>
>> >> I don't have the context, but if you are using Hadoop/Hbase, so don't
>> >> do
>> >> RAID on your disk.
>> >>
>> >>
>> >> On Fri, Feb 8, 2013 at 11:15 AM, Jean-Marc Spaggiari <
>> >> [EMAIL PROTECTED]> wrote:
>> >>
>> >> > Ok. I see. For my usecase I prefer to loose the data and have faster
>> >> > process. So I will go for RAID0 and keep the replication factor to
>> >> > 3... If at some point I have 5 disks in the node, I will most
>> >> > probably
>> >> > give a try to RAID5 and see the performances compared to the other
>> >> > RAID/JBOD options.
>> >> >
>> >> > Is there a "rule", like, 1 HD per core? Or we can't really simplify
>> >> > that
>> >> > much?
>> >> >
>> >> > So far I have that in the sar output:
>> >> > 21:35:03          tps      rtps      wtps   bread/s   bwrtn/s
>> >> > 21:45:03       218,85    215,97      2,88  45441,95    308,04
>> >> > 21:55:02       209,73    206,67      3,06  43985,28    378,32
>> >> > 22:05:04       215,03    211,71      3,33  44831,00    312,95
>> >> > Average :      214,54    211,45      3,09  44753,09    333,07
>> >> >
>> >> > But I'm not sure what it means. I will wait for tomorrow to get more
>> >> > results, but my job will be done over night, so I'm not sure the
>> >> > average will be accurate...
>> >> >
>> >> > JM
>> >> >
>> >> >
>> >> > 2013/2/7, Kevin O'dell <[EMAIL PROTECTED]>:
>> >> > > JM,
>> >> > >
>> >> > >   I think you misunderstood me.  I am not advocating any form of
>> RAID
>> >> for
>> >> > > Hadoop.  It is true that we already have redundancy built in with
>> >> > > HDFS.
>> >> >  So
>> >> > > unless you were going to do something silly like sacrifice speed
>> >> > > to
>> >> > > run
>> >> > > RAID1 or RAID5 and lower your replication to 2...just don't do it
>> >> > > :)
>> >> > >  Anyway, yes you probably should have 3 - 4 drives per node if not
>> >> more.
>> >> > >  At that point then the you will really see the benefit of JBOD