Hi,

I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
cores).

I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming jar
with 'wc -l' as mapper and 'cat' as reducer.

I use 64MB block size and the default replication (3).

The wc on the 100 GB took about 220 seconds which translates to about 3.5
Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec so
i would expect someting around 20 GBit/sec (minus some overhead), and I'm
getting only 3.5.

Is my expectaion valid?

I checked the jobtracked and it seems all nodes are working, each reading
the right blocks. I have not played with the number of mapper and reducers
yet. It seems the number of mappers is the same as the number of blocks and
the number of reducers is 20 (there are 20 disks). This looks ok for me.

We also did an experiment with TestDFSIO with similar results. Aggregated
read io speed is around 3.5Gbit/sec. It is just too far from my
expectation:(

Please help!

Thank you,
Gyorgy
--
View this message in context: http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
Sent from the Hadoop core-user mailing list archive at Nabble.com.

That's because you are assuming that processing time for mappers and reducers to be 0? Counting words is processor intensive and it's likely that lot of those 220 seconds are spent in processing, not just reading the file.

On May 30, 2011, at 8:28 AM, "ext Gyuribácsi" <[EMAIL PROTECTED]> wrote:

>
>
> Hi,
>
> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
> cores).
>
> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming jar
> with 'wc -l' as mapper and 'cat' as reducer.
>
> I use 64MB block size and the default replication (3).
>
> The wc on the 100 GB took about 220 seconds which translates to about 3.5
> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec so
> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
> getting only 3.5.
>
> Is my expectaion valid?
>
> I checked the jobtracked and it seems all nodes are working, each reading
> the right blocks. I have not played with the number of mapper and reducers
> yet. It seems the number of mappers is the same as the number of blocks and
> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>
> We also did an experiment with TestDFSIO with similar results. Aggregated
> read io speed is around 3.5Gbit/sec. It is just too far from my
> expectation:(
>
> Please help!
>
> Thank you,
> Gyorgy
> --
> View this message in context: http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
> Sent from the Hadoop core-user mailing list archive at Nabble.com.
>

On May 30, 2011, at 7:27 AM, Gyuribácsi wrote:

>
>
> Hi,
>
> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
> cores).
>
> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming jar
> with 'wc -l' as mapper and 'cat' as reducer.
>
> I use 64MB block size and the default replication (3).
>
> The wc on the 100 GB took about 220 seconds which translates to about 3.5
> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec so
> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
> getting only 3.5.
>
> Is my expectaion valid?

Probably not, at least not out-of-the-box.  Things to tune:
1) Number of threads per disk
2) Size of the block size (yours seems relatively small)
3) Network bottlenecks (can be seen using Ganglia)
4) Related to (3), the number of replicas.
5) Selection of the Linux I/O scheduler.  Default CFQ scheduler is inappropriate for batch workloads.

Finally, if you don't have enough host-level monitoring to indicate the current bottleneck (CPU, memory, network, or I/O?), you likely won't ever be able to solve this riddle

Brian

Ljddfjfjfififfifjftjiiiiiifjfjjjffkxbznzsjxodiewisshsudddudsjidhddueiweefiuftttoitfiirriifoiffkllddiririiriioerorooiieirrioeekroooeoooirjjfdijdkkduddjudiiehs
On May 30, 2011 5:28 AM, "Gyuribácsi" <[EMAIL PROTECTED]> wrote:
>
>
> Hi,
>
> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
> cores).
>
> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
jar
> with 'wc -l' as mapper and 'cat' as reducer.
>
> I use 64MB block size and the default replication (3).
>
> The wc on the 100 GB took about 220 seconds which translates to about 3.5
> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec
so
> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
> getting only 3.5.
>
> Is my expectaion valid?
>
> I checked the jobtracked and it seems all nodes are working, each reading
> the right blocks. I have not played with the number of mapper and reducers
> yet. It seems the number of mappers is the same as the number of blocks
and
> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>
> We also did an experiment with TestDFSIO with similar results. Aggregated
> read io speed is around 3.5Gbit/sec. It is just too far from my
> expectation:(
>
> Please help!
>
> Thank you,
> Gyorgy
> --
> View this message in context:
http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
> Sent from the Hadoop core-user mailing list archive at Nabble.com.
>

Not sure that will help ;)

Sent from my mobile. Please excuse the typos.

On 2011-05-30, at 9:23 AM, Boris Aleksandrovsky <[EMAIL PROTECTED]> wrote:

> Ljddfjfjfififfifjftjiiiiiifjfjjjffkxbznzsjxodiewisshsudddudsjidhddueiweefiuftttoitfiirriifoiffkllddiririiriioerorooiieirrioeekroooeoooirjjfdijdkkduddjudiiehs
> On May 30, 2011 5:28 AM, "Gyuribácsi" <[EMAIL PROTECTED]> wrote:
>>
>>
>> Hi,
>>
>> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
>> cores).
>>
>> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
> jar
>> with 'wc -l' as mapper and 'cat' as reducer.
>>
>> I use 64MB block size and the default replication (3).
>>
>> The wc on the 100 GB took about 220 seconds which translates to about 3.5
>> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec
> so
>> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
>> getting only 3.5.
>>
>> Is my expectaion valid?
>>
>> I checked the jobtracked and it seems all nodes are working, each reading
>> the right blocks. I have not played with the number of mapper and reducers
>> yet. It seems the number of mappers is the same as the number of blocks
> and
>> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>>
>> We also did an experiment with TestDFSIO with similar results. Aggregated
>> read io speed is around 3.5Gbit/sec. It is just too far from my
>> expectation:(
>>
>> Please help!
>>
>> Thank you,
>> Gyorgy
>> --
>> View this message in context:
> http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
>> Sent from the Hadoop core-user mailing list archive at Nabble.com.
>>

Psst. The cats speak in their own language ;-)

On Mon, May 30, 2011 at 10:31 PM, James Seigel <[EMAIL PROTECTED]> wrote:
> Not sure that will help ;)
>
> Sent from my mobile. Please excuse the typos.
>
> On 2011-05-30, at 9:23 AM, Boris Aleksandrovsky <[EMAIL PROTECTED]> wrote:
>
>> Ljddfjfjfififfifjftjiiiiiifjfjjjffkxbznzsjxodiewisshsudddudsjidhddueiweefiuftttoitfiirriifoiffkllddiririiriioerorooiieirrioeekroooeoooirjjfdijdkkduddjudiiehs
>> On May 30, 2011 5:28 AM, "Gyuribácsi" <[EMAIL PROTECTED]> wrote:
>>>
>>>
>>> Hi,
>>>
>>> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
>>> cores).
>>>
>>> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
>> jar
>>> with 'wc -l' as mapper and 'cat' as reducer.
>>>
>>> I use 64MB block size and the default replication (3).
>>>
>>> The wc on the 100 GB took about 220 seconds which translates to about 3.5
>>> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec
>> so
>>> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
>>> getting only 3.5.
>>>
>>> Is my expectaion valid?
>>>
>>> I checked the jobtracked and it seems all nodes are working, each reading
>>> the right blocks. I have not played with the number of mapper and reducers
>>> yet. It seems the number of mappers is the same as the number of blocks
>> and
>>> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>>>
>>> We also did an experiment with TestDFSIO with similar results. Aggregated
>>> read io speed is around 3.5Gbit/sec. It is just too far from my
>>> expectation:(
>>>
>>> Please help!
>>>
>>> Thank you,
>>> Gyorgy
>>> --
>>> View this message in context:
>> http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
>>> Sent from the Hadoop core-user mailing list archive at Nabble.com.
>>>
>

--
Harsh J

That's a small town in Iceland.

On Mon, May 30, 2011 at 10:01 AM, James Seigel <[EMAIL PROTECTED]> wrote:
> Not sure that will help ;)
>
> Sent from my mobile. Please excuse the typos.
>
> On 2011-05-30, at 9:23 AM, Boris Aleksandrovsky <[EMAIL PROTECTED]> wrote:
>
>> Ljddfjfjfififfifjftjiiiiiifjfjjjffkxbznzsjxodiewisshsudddudsjidhddueiweefiuftttoitfiirriifoiffkllddiririiriioerorooiieirrioeekroooeoooirjjfdijdkkduddjudiiehs
>> On May 30, 2011 5:28 AM, "Gyuribácsi" <[EMAIL PROTECTED]> wrote:
>>>
>>>
>>> Hi,
>>>
>>> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
>>> cores).
>>>
>>> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
>> jar
>>> with 'wc -l' as mapper and 'cat' as reducer.
>>>
>>> I use 64MB block size and the default replication (3).
>>>
>>> The wc on the 100 GB took about 220 seconds which translates to about 3.5
>>> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec
>> so
>>> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
>>> getting only 3.5.
>>>
>>> Is my expectaion valid?
>>>
>>> I checked the jobtracked and it seems all nodes are working, each reading
>>> the right blocks. I have not played with the number of mapper and reducers
>>> yet. It seems the number of mappers is the same as the number of blocks
>> and
>>> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>>>
>>> We also did an experiment with TestDFSIO with similar results. Aggregated
>>> read io speed is around 3.5Gbit/sec. It is just too far from my
>>> expectation:(
>>>
>>> Please help!
>>>
>>> Thank you,
>>> Gyorgy
>>> --
>>> View this message in context:
>> http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
>>> Sent from the Hadoop core-user mailing list archive at Nabble.com.
>>>
>

Hi Gyuribácsi

I would suggest you divide MapReduce program execution time into 3 parts

a) Map Stage
In this stage, wc splits input data and generates map tasks. Each map task
process one block (in default, you can change it in FileInputFormat.java).
As Brian said, if you have larger blocks size, you may have less number of
map tasks, and then probably less overhead.

b) Reduce Stage
    2) shuffle phase
     In this phase, reduce task collect intermediate results from every node
that has executed map tasks. Each reduce task can have many current threads
to obtain data(you can configure it in mapred-site.xml, it is
"mapreduce.reduce.shuffle.parallelcopies"). But, be careful to your data
popularity. For example, you have "Hadoop, Hadoop, Hadoop,hello". The
default Hadoop partitioner will assign 3 <Hadoop, 1>  key-value pairs to one
node. Thus, if you have two nodes run reduce tasks, one of them will copy 3
times ｍｏｒｅ data than the other. This will cause one node slower than the
other. You may rewrite the partitioner.

    3) sort and reduce phase
    I think the Hadoop UI ｗｉｌｌ ｇ���ｖｅ ｙｏｕ ｓｏ���ｅ ｈｉｎｔｓ ａ���ｏｕｔ ｈｏｗ ｌ���ｎｇ ｔｈｉｓ ｐ���ａｓｅ
ｔａｋｅｓ．

Ｂｙ ｄｉｖｉｄｉｎｇ ＭａｐＲｅｄｕｃｅ ａｐｐｌｉｃａｔｉｏｎ ｉｎｔｏ ｔｈｅｓｅ ３ ｐａｒｔｓ， ｙｏｕ ｃａｎ ｅａｓｉｌｙ ｆｉｎｄ
ｗｈｉｃｈ ｏｎｅ ｉｓ ｙｏｕｒ ｂｏｔｔｌｅｎｅｃｋ ａｎｄ ｄｏ ｓｏｍｅ ｐｒｏｆｉｌｉｎｇ． Ａｎｄ Ｉ ｄｏｎ＇ｔ ｋｎｏｗ ｗｈｙ ｍｙ
ｆｏｎｔ ｃｈａｎｇｅ ｔｏ ｔｈｉｓ ｔｙｐｅ．：（

Ｈｏｐｅ ｉｔ ｗｉｌｌ ｂｅ ｈｅｌｐｆｕｌ．
Ｃｈｅｎ

On Mon, May 30, 2011 at 12:32 PM, Harsh J <[EMAIL PROTECTED]> wrote:

> Psst. The cats speak in their own language ;-)
>
> On Mon, May 30, 2011 at 10:31 PM, James Seigel <[EMAIL PROTECTED]> wrote:
> > Not sure that will help ;)
> >
> > Sent from my mobile. Please excuse the typos.
> >
> > On 2011-05-30, at 9:23 AM, Boris Aleksandrovsky <[EMAIL PROTECTED]>
> wrote:
> >
> >>
> Ljddfjfjfififfifjftjiiiiiifjfjjjffkxbznzsjxodiewisshsudddudsjidhddueiweefiuftttoitfiirriifoiffkllddiririiriioerorooiieirrioeekroooeoooirjjfdijdkkduddjudiiehs
> >> On May 30, 2011 5:28 AM, "Gyuribácsi" <[EMAIL PROTECTED]> wrote:
> >>>
> >>>
> >>> Hi,
> >>>
> >>> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16
> HT
> >>> cores).
> >>>
> >>> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
> >> jar
> >>> with 'wc -l' as mapper and 'cat' as reducer.
> >>>
> >>> I use 64MB block size and the default replication (3).
> >>>
> >>> The wc on the 100 GB took about 220 seconds which translates to about
> 3.5
> >>> Gbit/sec processing speed. One disk can do sequential read with
> 1Gbit/sec
> >> so
> >>> i would expect someting around 20 GBit/sec (minus some overhead), and
> I'm
> >>> getting only 3.5.
> >>>
> >>> Is my expectaion valid?
> >>>
> >>> I checked the jobtracked and it seems all nodes are working, each
> reading
> >>> the right blocks. I have not played with the number of mapper and
> reducers
> >>> yet. It seems the number of mappers is the same as the number of blocks
> >> and
> >>> the number of reducers is 20 (there are 20 disks). This looks ok for
> me.
> >>>
> >>> We also did an experiment with TestDFSIO with similar results.
> Aggregated
> >>> read io speed is around 3.5Gbit/sec. It is just too far from my
> >>> expectation:(
> >>>
> >>> Please help!
> >>>
> >>> Thank you,
> >>> Gyorgy
> >>> --
> >>> View this message in context:
> >>
> http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
> >>> Sent from the Hadoop core-user mailing list archive at Nabble.com.
> >>>
> >
>
>
>
> --
> Harsh J
>

Your Font block size got increased dynamically , check in core-site :) :)

- Jagaran

________________________________
From: He Chen <[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Sent: Mon, 30 May, 2011 11:39:35 AM
Subject: Re: Poor IO performance on a 10 node cluster.

Hi Gyuribácsi

I would suggest you divide MapReduce program execution time into 3 parts

a) Map Stage
In this stage, wc splits input data and generates map tasks. Each map task
process one block (in default, you can change it in FileInputFormat.java).
As Brian said, if you have larger blocks size, you may have less number of
map tasks, and then probably less overhead.

b) Reduce Stage
    2) shuffle phase
     In this phase, reduce task collect intermediate results from every node
that has executed map tasks. Each reduce task can have many current threads
to obtain data(you can configure it in mapred-site.xml, it is
"mapreduce.reduce.shuffle.parallelcopies"). But, be careful to your data
popularity. For example, you have "Hadoop, Hadoop, Hadoop,hello". The
default Hadoop partitioner will assign 3 <Hadoop, 1>  key-value pairs to one
node. Thus, if you have two nodes run reduce tasks, one of them will copy 3
times ｍｏｒｅ data than the other. This will cause one node slower than the
other. You may rewrite the partitioner.

    3) sort and reduce phase
    I think the Hadoop UI ｗｉｌｌ ｇｉｖｅ ｙｏｕ ｓｏｍｅ ｈｉｎｔｓ ａｂｏｕｔ ｈｏｗ ｌｏｎｇ ｔｈｉｓ ｐｈａｓｅ
��ａｋｅｓ．

Ｂｙ ｄｉｖｉｄｉｎｇ ＭａｐＲｅｄｕｃ��� ａｐｐｌｉｃａ���ｉｏｎ ｉｎｔｏ ���ｈｅｓｅ ３ ｐａ���ｔｓ， ｙｏｕ ｃ���ｎ ｅａｓｉｌｙ ���ｉｎｄ
ｗｈｉｃ��� ｏｎｅ ｉｓ ｙｏ���ｒ ｂｏｔｔｌｅ���ｅｃｋ ａｎｄ ｄ��� ｓｏｍｅ ｐｒｏ���ｉｌｉｎｇ． Ａ���ｄ Ｉ ｄｏｎ＇ｔ ���ｎｏｗ ｗｈｙ ｍ���
ｆｏｎｔ ｃｈａｎｇｅ ｔｏ ｔｈｉｓ ｔｙｐｅ．：（

Ｈｏｐｅ ｉｔ ｗ���ｌｌ ｂｅ ｈｅｌ���ｆｕｌ．
Ｃｈｅ���

On Mon, May 30, 2011 at 12:32 PM, Harsh J <[EMAIL PROTECTED]> wrote:

> Psst. The cats speak in their own language ;-)
>
> On Mon, May 30, 2011 at 10:31 PM, James Seigel <[EMAIL PROTECTED]> wrote:
> > Not sure that will help ;)
> >
> > Sent from my mobile. Please excuse the typos.
> >
> > On 2011-05-30, at 9:23 AM, Boris Aleksandrovsky <[EMAIL PROTECTED]>
> wrote:
> >
> >>
>Ljddfjfjfififfifjftjiiiiiifjfjjjffkxbznzsjxodiewisshsudddudsjidhddueiweefiuftttoitfiirriifoiffkllddiririiriioerorooiieirrioeekroooeoooirjjfdijdkkduddjudiiehs
>s
> >> On May 30, 2011 5:28 AM, "Gyuribácsi" <[EMAIL PROTECTED]> wrote:
> >>>
> >>>
> >>> Hi,
> >>>
> >>> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16
> HT
> >>> cores).
> >>>
> >>> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
> >> jar
> >>> with 'wc -l' as mapper and 'cat' as reducer.
> >>>
> >>> I use 64MB block size and the default replication (3).
> >>>
> >>> The wc on the 100 GB took about 220 seconds which translates to about
> 3.5
> >>> Gbit/sec processing speed. One disk can do sequential read with
> 1Gbit/sec
> >> so
> >>> i would expect someting around 20 GBit/sec (minus some overhead), and
> I'm
> >>> getting only 3.5.
> >>>
> >>> Is my expectaion valid?
> >>>
> >>> I checked the jobtracked and it seems all nodes are working, each
> reading
> >>> the right blocks. I have not played with the number of mapper and
> reducers
> >>> yet. It seems the number of mappers is the same as the number of blocks
> >> and
> >>> the number of reducers is 20 (there are 20 disks). This looks ok for
> me.
> >>>
> >>> We also did an experiment with TestDFSIO with similar results.
> Aggregated
> >>> read io speed is around 3.5Gbit/sec. It is just too far from my
> >>> expectation:(
> >>>
> >>> Please help!
> >>>
> >>> Thank you,
> >>> Gyorgy
> >>> --
> >>> View this message in context:
> >>
>http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
>l
> >>> Sent from the Hadoop core-user mailing list archive at Nabble.com.
> >>>
> >
>
>
>
> --
> Harsh J
>

I'm sorry, but she's with me now.

On Mon, May 30, 2011 at 8:22 AM, Boris Aleksandrovsky
<[EMAIL PROTECTED]> wrote:
> Ljddfjfjfififfifjftjiiiiiifjfjjjffkxbznzsjxodiewisshsudddudsjidhddueiweefiuftttoitfiirriifoiffkllddiririiriioerorooiieirrioeekroooeoooirjjfdijdkkduddjudiiehs
> On May 30, 2011 5:28 AM, "Gyuribácsi" <[EMAIL PROTECTED]> wrote:
>>
>>
>> Hi,
>>
>> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
>> cores).
>>
>> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
> jar
>> with 'wc -l' as mapper and 'cat' as reducer.
>>
>> I use 64MB block size and the default replication (3).
>>
>> The wc on the 100 GB took about 220 seconds which translates to about 3.5
>> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec
> so
>> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
>> getting only 3.5.
>>
>> Is my expectaion valid?
>>
>> I checked the jobtracked and it seems all nodes are working, each reading
>> the right blocks. I have not played with the number of mapper and reducers
>> yet. It seems the number of mappers is the same as the number of blocks
> and
>> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>>
>> We also did an experiment with TestDFSIO with similar results. Aggregated
>> read io speed is around 3.5Gbit/sec. It is just too far from my
>> expectation:(
>>
>> Please help!
>>
>> Thank you,
>> Gyorgy
>> --
>> View this message in context:
> http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p31732971.html
>> Sent from the Hadoop core-user mailing list archive at Nabble.com.
>>
>

--
Lance Norskog
[EMAIL PROTECTED]

Some things which helped us include setting your vm.swappiness to 0 and
mounting your disks with noatime,nodiratime options.

Also make sure your disks aren't setup with RAID (JBOD is recommended)

You might want to run terasort as you tweak your environment.  It's very
helpful when checking if a change helped (or hurt) your cluster.

Hope that helps a bit.

On 05/30/2011 06:27 AM, Gyurib�csi wrote:
>
> Hi,
>
> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
> cores).
>
> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming jar
> with 'wc -l' as mapper and 'cat' as reducer.
>
> I use 64MB block size and the default replication (3).
>
> The wc on the 100 GB took about 220 seconds which translates to about 3.5
> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec so
> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
> getting only 3.5.
>
> Is my expectaion valid?
>
> I checked the jobtracked and it seems all nodes are working, each reading
> the right blocks. I have not played with the number of mapper and reducers
> yet. It seems the number of mappers is the same as the number of blocks and
> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>
> We also did an experiment with TestDFSIO with similar results. Aggregated
> read io speed is around 3.5Gbit/sec. It is just too far from my
> expectation:(
>
> Please help!
>
> Thank you,
> Gyorgy
>

It is also worth using dd to verify your raw disk speeds.

Also, expressing disk transfer rates in bytes per second makes it a bit
easier for most of the disk people I know to figure out what is large or
small.

Each of these disks disk should do about 100MB/s when driven well.  Hadoop
does OK, but not nearly full capacity so I would expected more like
40-50MB/s/disk.  Also, if one of your disks is doing double duty you may
have some extra cost.  40 x 2 x 10 = 800MB per second.  You are doing 100GB
/ 220 seconds = 0.5 GB / s which isn't so terribly bad.  It is definitely
less than the theoretically possible 100 x 2 x 10 = 2GB/second and I would
expect you could tune this up a little, but not a massive amount.

In general, blade servers do not make good Hadoop nodes exactly because the
I/O performance tends to be low when you only have a few spindle.

One other reason that this might be a bit below expectations is that your
files may not be well distributed on your cluster.  Can you say what you
used to upload the files?

On Wed, Jun 1, 2011 at 5:56 PM, hadoopman <[EMAIL PROTECTED]> wrote:

> Some things which helped us include setting your vm.swappiness to 0 and
> mounting your disks with noatime,nodiratime options.
>
> Also make sure your disks aren't setup with RAID (JBOD is recommended)
>
> You might want to run terasort as you tweak your environment.  It's very
> helpful when checking if a change helped (or hurt) your cluster.
>
> Hope that helps a bit.
>
> On 05/30/2011 06:27 AM, Gyuribácsi wrote:
>
>>
>> Hi,
>>
>> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
>> cores).
>>
>> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
>> jar
>> with 'wc -l' as mapper and 'cat' as reducer.
>>
>> I use 64MB block size and the default replication (3).
>>
>> The wc on the 100 GB took about 220 seconds which translates to about 3.5
>> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec
>> so
>> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
>> getting only 3.5.
>>
>> Is my expectaion valid?
>>
>> I checked the jobtracked and it seems all nodes are working, each reading
>> the right blocks. I have not played with the number of mapper and reducers
>> yet. It seems the number of mappers is the same as the number of blocks
>> and
>> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>>
>> We also did an experiment with TestDFSIO with similar results. Aggregated
>> read io speed is around 3.5Gbit/sec. It is just too far from my
>> expectation:(
>>
>> Please help!
>>
>> Thank you,
>> Gyorgy
>>
>>
>
>

Hi,

Is this the speed you are observing when initial writes of the files
happening (i.e while you are initially putting 10gb files with replication)

B.T.W can you check the speed with one datanode and no replication..

The below URL would have some more combinations you could try
http://lucene.472066.n3.nabble.com/HDFS-block-size-v-s-mapred-min-split-size-td2521091.html

Regards,
Raja Nagendra Kumar
Gyuribácsi wrote:
>
>  
> Hi,
>
> I have a 10 node cluster (IBM blade servers, 48GB RAM, 2x500GB Disk, 16 HT
> cores).
>
> I've uploaded 10 files to HDFS. Each file is 10GB. I used the streaming
> jar
> with 'wc -l' as mapper and 'cat' as reducer.
>
> I use 64MB block size and the default replication (3).
>
> The wc on the 100 GB took about 220 seconds which translates to about 3.5
> Gbit/sec processing speed. One disk can do sequential read with 1Gbit/sec
> so
> i would expect someting around 20 GBit/sec (minus some overhead), and I'm
> getting only 3.5.
>
> Is my expectaion valid?
>
> I checked the jobtracked and it seems all nodes are working, each reading
> the right blocks. I have not played with the number of mapper and reducers
> yet. It seems the number of mappers is the same as the number of blocks
> and
> the number of reducers is 20 (there are 20 disks). This looks ok for me.
>
> We also did an experiment with TestDFSIO with similar results. Aggregated
> read io speed is around 3.5Gbit/sec. It is just too far from my
> expectation:(
>
> Please help!
>
> Thank you,
> Gyorgy
>

--
View this message in context: http://old.nabble.com/Poor-IO-performance-on-a-10-node-cluster.-tp31732971p32076106.html
Sent from the Hadoop core-user mailing list archive at Nabble.com.