Hi,
I have some quetions about framework scheduler and DRF Allocator.
When I launch the mesos-master, I set --weights=hadoop=1,torque=2.
But, even if torque has no job to run , the Master still give the 2/3 resources
of the whole system to it. And these resources cann't be used by hadoop.
Is ther
Dear all,
I wrote a framework scheduler with c++. But I don't know how compile it.
please tell me how to compile my own framework scheduler and executor.
PS: why the long_lived_framework.cpp is included in Makefile, but after running
the make command, there is no long_lived_framework.o or long
Thank you very much !
Best Wishes !
在2013-11-26 02:43:50,HUO Jing写道:
> After running make, the long_lived_framework binary is placed in the src
> directory of your build directory. However, this is a binary generated by
> libtool.
>
> You will need to ensure the mesos headers
Thank you very mach !
在2013-11-26 00:34:14,HUO Jing写道:
> Hi Huo,
>
> If you have run `make install`, your library and include path just need to
> contain /usr/local/lib and /usr/local/include respectively.
> You can check your installation by looking for /usr/local/include/me
Hi,
I have a question about resourceOffer.
Now an offer presented to the framework will not be rescinded until
launchTask() is called. If framework1 has no Task for it, how can I pass this
offer to another framework?
Can I use offerRescinded() or declineOffer()?
Thank you very much
Huojin from
Hi,
I am trying to develop a framework scheduler and executor to add Torque to
Mesos.
Now, I have a problem. As I know, Mesos uses linux container to isolate
resources, so pbs_mom has to be started in executor. But when there is no task
running in the executor, mesos slave will shut down the exe
Thank you very much!
Another question:
Can I start the Executor even if there is no task to run on it?
And how to do it ?
> -原始邮件-
> 发件人: "Vinod Kone"
> 发送时间: 2014年1月6日 星期一
> 收件人: dev
> 抄送:
> 主题: Re: Question about Executor
>
> On Sun, Jan 5
e it.
> >
> > https://github.com/apache/mesos/tree/master/frameworks/torque
> >
> > Enjoy!
> > Julien
> >
> >
> > On Mon, Jan 6, 2014 at 4:13 AM, HUO Jing wrote:
> >
> > > Thank you very much!
> > > Another question:
> >
Hi,
Is there any other Resource allocation algorithm for mesos ?
The DRF algorithm is not very suitable for me.
I want to use an allocation algorithm which is like the capacity scheduler of
Hadoop.
Thank you very much!
Huojing
Hi,
I meet a problem during make the mesos-0.14.0
please tell me how to solve this problem
*** Warning: Linking the shared library libmesos.la against the
*** static library ../3rdparty/leveldb/libleveldb.a is not portable!
libtool: link: g++ -fPIC -DPIC -shared -nostdlib
/usr/lib/gcc/x86_64-re
Hi,
I have installed Mesos and Hadoop CDH4.5.0, changed the mapred-site.xml, and
packaged hadoop-mesos-0.0.5.jar with hadoop and upload it to hdfs. In a word, I
have done everything in this page: https://github.com/mesos/hadoop.
but when I try to run jobtracker with command:
bash-3.2$ /usr/local/
it in other ways instead
> (in that user's .bashrc or by creating a wrapper around 'hadoop' that sets
> the variable before launching 'hadoop').
> You are very close to having Hadoop running on top of Mesos.
> Good luck!
> -Adam-
>
>
> On Tue, Jan
we set that environment
> variable on the command-line first. You could set it in other ways instead
> (in that user's .bashrc or by creating a wrapper around 'hadoop' that sets
> the variable before launching 'hadoop').
> You are very close to having Hadoop r
sorry, this is my fault, there are two diffenrent version in my host.
> -原始邮件-
> 发件人: "HUO Jing"
> 发送时间: 2014年1月15日 星期三
> 收件人: dev@mesos.apache.org
> 抄送: mesos-dev
> 主题: Re: Re: How to run hadoop Jobtracker
>
> When I try this
> command"ME
you setting the config in?
> 14/01/15 13:09:15 INFO ipc.Server: Starting Socket Reader #1 for port 8021/
> 14/01/15 13:09:15 INFO metrics.RpcMetrics: Initializing RPC Metrics with
> hostName=JobTracker, port=8021
>
>
> On Tue, Jan 14, 2014 at 9:29 PM, HUO Jing wrote:
>
>
;start' action is doing behind the scenes.
> Chances are it's updating a library path and other environment settings
> that you won't get just from running 'hadoop jobtracker'
>
>
> On Tue, Jan 14, 2014 at 10:59 PM, HUO Jing wrote:
>
> > I
shell script.
> Look inside and see what the 'start' action is doing behind the scenes.
> Chances are it's updating a library path and other environment settings
> that you won't get just from running 'hadoop jobtracker'
>
>
> On Tue, Jan 14, 2014
7; action is doing behind the scenes.
> Chances are it's updating a library path and other environment settings
> that you won't get just from running 'hadoop jobtracker'
>
>
> On Tue, Jan 14, 2014 at 10:59 PM, HUO Jing wrote:
>
> > I find out that even w
Hi,
I am trying to run Hadoop on Mesos, but the stderr shows:
Exception in thread "main" java.lang.NoClassDefFoundError:
org/apache/hadoop/mapred/MesosExecutor
Caused by: java.lang.ClassNotFoundException:
org.apache.hadoop.mapred.MesosExecutor
at java.net.URLClassLoader$1.run(URLClassLoa
on Mesos.
On Fri, Jan 17, 2014 at 10:09 AM, HUO Jing wrote:
> Hi,
> I am trying to run Hadoop on Mesos, but the stderr shows:
>
> Exception in thread "main" java.lang.NoClassDefFoundError:
> org/apache/hadoop/mapred/MesosExecutor
> Caused by: java.lang.ClassNotFoundE
Yes, I followed the instructions.
I put the hadoop-mesos.jar in hadoop-2.0.0-mr1-cdh4.2.2/lib/ and tar it to
hadoop-2.0.0-mr1-cdh4.2.2.tar.gz
then upload it to HDFS.
-原始邮件-
发件人: "Brenden Matthews"
发送时间: 2014年1月24日 星期五
收件人: "HUO Jing"
抄送: dev
主题: Re: F
Hi,
I ran a map-only job on hadoop. But on every tasktracker, it had only 7 map
slots running. How to let it be 8 map slots?
Hi,
I ran a map-only job on hadoop. Every node has 8 cores, but on every
tasktracker, it had only 7 map slots running. How to let tasktracker run 8 map
slots on every node?
Thank you
Another question,
The tasktracker can not stop after the hadoop job is done. Why?
> -原始邮件-
> 发件人: "HUO Jing"
> 发送时间: 2014年1月27日 星期一
> 收件人: dev@mesos.apache.org
> 抄送: "Brenden Matthews"
> 主题: Re: question about hadoop on mesos
>
> Hi,
> I
Hi,
I have some questions about running hadoop on top of Mesos, please help me.
1. when a tasktracker is launched, if n cpu core are allocated to it, it can
only launch n-1 map tasks. Could someone tell me why? And, if I want to run
map-only job, what should I do to run n map tasks on a n cpu res
> 2. After a tasktracker is launched, in what condition it's status will
> > update to FINISHED?
> > In my cluster, sometimes it will never end until I restart the jobtracker.
> > Sometimes it will end if there is no task or job in jobtracker to run.
>
> The expected case is that the task tracker
t; 发送时间: 2014年1月28日 星期二
> 收件人: dev
> 抄送: "u...@mesos.apache.org"
> 主题: Re: Re: Please Help me about hadoop on Mesos
>
> On Mon, Jan 27, 2014 at 10:07 AM, HUO Jing wrote:
>
> > So, at the very beginning, if all the resource are assigned to hadoop, and
> >
Hi,
I have some questions about the DRF allocation.
1, How to set weight for Role?
2, If one Role have several frameworks, can we set different weights for every
framework? And How to do it?
3, I have read the code of drf_sorter.cpp and drf_sorter.hpp and
hierarchical_allocator_process.hpp,
In Ro
Hi,
When I try to start master with set a role and its weight, there is an error.
[root@hadoop06 qsub]# mesos-master --roles=role1 --weights='role1=2'
I0228 23:23:28.682869 744 main.cpp:114] Build: 2014-01-13 21:30:01 by root
I0228 23:23:28.683229 744 main.cpp:115] Starting Mesos master
I0228
ights when starting the Mesos Master
>
> On Fri, Feb 28, 2014 at 7:32 AM, HUO Jing wrote:
>
> > mesos-master --roles=role1 --weights='role1=2'
> >
>
> I'm unable to reproduce this? Are you sure that is the comand line you
> used? What version of
Hi,
I met an error while installing mesos. When I run 'make' command, it shows:
make[2]: ../3rdparty/libprocess/3rdparty/protobuf-2.4.1/src/protoc: Command not
found
make[2]: ***
[../3rdparty/libprocess/3rdparty/protobuf-2.4.1/src/google/protobuf/descriptor.pb.h]
Error 127
make[2]: Leaving dire
31 matches
Mail list logo