You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@zookeeper.apache.org by vilobh meshram <me...@gmail.com> on 2011/06/20 23:36:30 UTC
Maximum Read/Write throughput for Zookeeper
Hi,
I am working on a project where we are making use of Zookeeper.
I wanted to know the maximum throughput Zookeeper can provide for read/write
workloads.
Also if there are any standard benchmarks which can be used to measure such
throughput can be useful.
Also if you can point me to some paper or documents which mention these
details will be useful.
Thanks,
Vilobh
Graduate Research Associate
Department of Computer Science
The Ohio State University Columbus Ohio
Re: Maximum Read/Write throughput for Zookeeper
Posted by Ted Dunning <te...@gmail.com>.
As a pound of flesh in return, could I ask you to post your results on an
Apache wiki?
On Tue, Jun 21, 2011 at 12:04 AM, vilobh meshram
<me...@gmail.com>wrote:
> Thanks Ted.
>
>
> On Mon, Jun 20, 2011 at 5:56 PM, Ted Dunning <te...@gmail.com>wrote:
>
>> At the risk of being rude, a quick search on Google for [zookeeper
>> performance] gives the following top hit:
>>
>> http://wiki.apache.org/hadoop/ZooKeeper/Performance
>>
>> The very nice graph that Patrick did many moons ago is missing, but the
>> link to
>>
>> http://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview
>>
>> provides some very nice information.
>>
>> I hate it when people give snarky answers about doing your homework before
>> asking in public, but it would be very easy to respond rudely to your
>> request. The Zookeeper mailing list is full of friendly people so I won't
>> say much more than that.
>>
>> In general, when sending requests to mailing lists, you will receive much
>> more helpful answers if you demonstrate that you have done a bit of work
>> ahead of time instead of depending on the mailing list as your first
>> resource. This link has some good suggestions:
>> http://www.catb.org/~esr/faqs/smart-questions.html
>>
>> Making sure that you do this is very important on mailing lists where
>> there is less of a premium placed on politeness than there is on most Apache
>> mailing lists. In some mailing lists, asking a poorly researched question
>> will subject you to being roasted alive or ignored completely. This is a
>> sad fact of life and it turns lots of newcomers off, but it absolutely does
>> happen.
>>
>>
>> On Mon, Jun 20, 2011 at 11:36 PM, vilobh meshram <
>> meshram.vilobh@gmail.com> wrote:
>>
>>> Hi,
>>>
>>> I am working on a project where we are making use of Zookeeper.
>>>
>>> I wanted to know the maximum throughput Zookeeper can provide for
>>> read/write
>>> workloads.
>>>
>>> Also if there are any standard benchmarks which can be used to measure
>>> such
>>> throughput can be useful.
>>>
>>> Also if you can point me to some paper or documents which mention these
>>> details will be useful.
>>>
>>> Thanks,
>>> Vilobh
>>> Graduate Research Associate
>>> Department of Computer Science
>>> The Ohio State University Columbus Ohio
>>>
>>
>>
>
Re: Maximum Read/Write throughput for Zookeeper
Posted by Ted Dunning <te...@gmail.com>.
As a pound of flesh in return, could I ask you to post your results on an
Apache wiki?
On Tue, Jun 21, 2011 at 12:04 AM, vilobh meshram
<me...@gmail.com>wrote:
> Thanks Ted.
>
>
> On Mon, Jun 20, 2011 at 5:56 PM, Ted Dunning <te...@gmail.com>wrote:
>
>> At the risk of being rude, a quick search on Google for [zookeeper
>> performance] gives the following top hit:
>>
>> http://wiki.apache.org/hadoop/ZooKeeper/Performance
>>
>> The very nice graph that Patrick did many moons ago is missing, but the
>> link to
>>
>> http://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview
>>
>> provides some very nice information.
>>
>> I hate it when people give snarky answers about doing your homework before
>> asking in public, but it would be very easy to respond rudely to your
>> request. The Zookeeper mailing list is full of friendly people so I won't
>> say much more than that.
>>
>> In general, when sending requests to mailing lists, you will receive much
>> more helpful answers if you demonstrate that you have done a bit of work
>> ahead of time instead of depending on the mailing list as your first
>> resource. This link has some good suggestions:
>> http://www.catb.org/~esr/faqs/smart-questions.html
>>
>> Making sure that you do this is very important on mailing lists where
>> there is less of a premium placed on politeness than there is on most Apache
>> mailing lists. In some mailing lists, asking a poorly researched question
>> will subject you to being roasted alive or ignored completely. This is a
>> sad fact of life and it turns lots of newcomers off, but it absolutely does
>> happen.
>>
>>
>> On Mon, Jun 20, 2011 at 11:36 PM, vilobh meshram <
>> meshram.vilobh@gmail.com> wrote:
>>
>>> Hi,
>>>
>>> I am working on a project where we are making use of Zookeeper.
>>>
>>> I wanted to know the maximum throughput Zookeeper can provide for
>>> read/write
>>> workloads.
>>>
>>> Also if there are any standard benchmarks which can be used to measure
>>> such
>>> throughput can be useful.
>>>
>>> Also if you can point me to some paper or documents which mention these
>>> details will be useful.
>>>
>>> Thanks,
>>> Vilobh
>>> Graduate Research Associate
>>> Department of Computer Science
>>> The Ohio State University Columbus Ohio
>>>
>>
>>
>
Re: Maximum Read/Write throughput for Zookeeper
Posted by vilobh meshram <me...@gmail.com>.
Thanks Ted.
On Mon, Jun 20, 2011 at 5:56 PM, Ted Dunning <te...@gmail.com> wrote:
> At the risk of being rude, a quick search on Google for [zookeeper
> performance] gives the following top hit:
>
> http://wiki.apache.org/hadoop/ZooKeeper/Performance
>
> The very nice graph that Patrick did many moons ago is missing, but the
> link to
>
> http://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview
>
> provides some very nice information.
>
> I hate it when people give snarky answers about doing your homework before
> asking in public, but it would be very easy to respond rudely to your
> request. The Zookeeper mailing list is full of friendly people so I won't
> say much more than that.
>
> In general, when sending requests to mailing lists, you will receive much
> more helpful answers if you demonstrate that you have done a bit of work
> ahead of time instead of depending on the mailing list as your first
> resource. This link has some good suggestions:
> http://www.catb.org/~esr/faqs/smart-questions.html
>
> Making sure that you do this is very important on mailing lists where there
> is less of a premium placed on politeness than there is on most Apache
> mailing lists. In some mailing lists, asking a poorly researched question
> will subject you to being roasted alive or ignored completely. This is a
> sad fact of life and it turns lots of newcomers off, but it absolutely does
> happen.
>
>
> On Mon, Jun 20, 2011 at 11:36 PM, vilobh meshram <meshram.vilobh@gmail.com
> > wrote:
>
>> Hi,
>>
>> I am working on a project where we are making use of Zookeeper.
>>
>> I wanted to know the maximum throughput Zookeeper can provide for
>> read/write
>> workloads.
>>
>> Also if there are any standard benchmarks which can be used to measure
>> such
>> throughput can be useful.
>>
>> Also if you can point me to some paper or documents which mention these
>> details will be useful.
>>
>> Thanks,
>> Vilobh
>> Graduate Research Associate
>> Department of Computer Science
>> The Ohio State University Columbus Ohio
>>
>
>
Re: Maximum Read/Write throughput for Zookeeper
Posted by vilobh meshram <me...@gmail.com>.
Thanks Ted.
On Mon, Jun 20, 2011 at 5:56 PM, Ted Dunning <te...@gmail.com> wrote:
> At the risk of being rude, a quick search on Google for [zookeeper
> performance] gives the following top hit:
>
> http://wiki.apache.org/hadoop/ZooKeeper/Performance
>
> The very nice graph that Patrick did many moons ago is missing, but the
> link to
>
> http://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview
>
> provides some very nice information.
>
> I hate it when people give snarky answers about doing your homework before
> asking in public, but it would be very easy to respond rudely to your
> request. The Zookeeper mailing list is full of friendly people so I won't
> say much more than that.
>
> In general, when sending requests to mailing lists, you will receive much
> more helpful answers if you demonstrate that you have done a bit of work
> ahead of time instead of depending on the mailing list as your first
> resource. This link has some good suggestions:
> http://www.catb.org/~esr/faqs/smart-questions.html
>
> Making sure that you do this is very important on mailing lists where there
> is less of a premium placed on politeness than there is on most Apache
> mailing lists. In some mailing lists, asking a poorly researched question
> will subject you to being roasted alive or ignored completely. This is a
> sad fact of life and it turns lots of newcomers off, but it absolutely does
> happen.
>
>
> On Mon, Jun 20, 2011 at 11:36 PM, vilobh meshram <meshram.vilobh@gmail.com
> > wrote:
>
>> Hi,
>>
>> I am working on a project where we are making use of Zookeeper.
>>
>> I wanted to know the maximum throughput Zookeeper can provide for
>> read/write
>> workloads.
>>
>> Also if there are any standard benchmarks which can be used to measure
>> such
>> throughput can be useful.
>>
>> Also if you can point me to some paper or documents which mention these
>> details will be useful.
>>
>> Thanks,
>> Vilobh
>> Graduate Research Associate
>> Department of Computer Science
>> The Ohio State University Columbus Ohio
>>
>
>
Re: Maximum Read/Write throughput for Zookeeper
Posted by Ted Dunning <te...@gmail.com>.
At the risk of being rude, a quick search on Google for [zookeeper
performance] gives the following top hit:
http://wiki.apache.org/hadoop/ZooKeeper/Performance
The very nice graph that Patrick did many moons ago is missing, but the link
to
http://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview
provides some very nice information.
I hate it when people give snarky answers about doing your homework before
asking in public, but it would be very easy to respond rudely to your
request. The Zookeeper mailing list is full of friendly people so I won't
say much more than that.
In general, when sending requests to mailing lists, you will receive much
more helpful answers if you demonstrate that you have done a bit of work
ahead of time instead of depending on the mailing list as your first
resource. This link has some good suggestions:
http://www.catb.org/~esr/faqs/smart-questions.html
Making sure that you do this is very important on mailing lists where there
is less of a premium placed on politeness than there is on most Apache
mailing lists. In some mailing lists, asking a poorly researched question
will subject you to being roasted alive or ignored completely. This is a
sad fact of life and it turns lots of newcomers off, but it absolutely does
happen.
On Mon, Jun 20, 2011 at 11:36 PM, vilobh meshram
<me...@gmail.com>wrote:
> Hi,
>
> I am working on a project where we are making use of Zookeeper.
>
> I wanted to know the maximum throughput Zookeeper can provide for
> read/write
> workloads.
>
> Also if there are any standard benchmarks which can be used to measure such
> throughput can be useful.
>
> Also if you can point me to some paper or documents which mention these
> details will be useful.
>
> Thanks,
> Vilobh
> Graduate Research Associate
> Department of Computer Science
> The Ohio State University Columbus Ohio
>
Re: Maximum Read/Write throughput for Zookeeper
Posted by Ted Dunning <te...@gmail.com>.
At the risk of being rude, a quick search on Google for [zookeeper
performance] gives the following top hit:
http://wiki.apache.org/hadoop/ZooKeeper/Performance
The very nice graph that Patrick did many moons ago is missing, but the link
to
http://wiki.apache.org/hadoop/ZooKeeper/ServiceLatencyOverview
provides some very nice information.
I hate it when people give snarky answers about doing your homework before
asking in public, but it would be very easy to respond rudely to your
request. The Zookeeper mailing list is full of friendly people so I won't
say much more than that.
In general, when sending requests to mailing lists, you will receive much
more helpful answers if you demonstrate that you have done a bit of work
ahead of time instead of depending on the mailing list as your first
resource. This link has some good suggestions:
http://www.catb.org/~esr/faqs/smart-questions.html
Making sure that you do this is very important on mailing lists where there
is less of a premium placed on politeness than there is on most Apache
mailing lists. In some mailing lists, asking a poorly researched question
will subject you to being roasted alive or ignored completely. This is a
sad fact of life and it turns lots of newcomers off, but it absolutely does
happen.
On Mon, Jun 20, 2011 at 11:36 PM, vilobh meshram
<me...@gmail.com>wrote:
> Hi,
>
> I am working on a project where we are making use of Zookeeper.
>
> I wanted to know the maximum throughput Zookeeper can provide for
> read/write
> workloads.
>
> Also if there are any standard benchmarks which can be used to measure such
> throughput can be useful.
>
> Also if you can point me to some paper or documents which mention these
> details will be useful.
>
> Thanks,
> Vilobh
> Graduate Research Associate
> Department of Computer Science
> The Ohio State University Columbus Ohio
>