You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Jim Donahue <jd...@adobe.com> on 2012/06/06 00:39:19 UTC
Problem running Mesos tests
I've been trying to build an EC2 instance that has Mesos installed (along with some other software). I started with the "standard" Amazon Linux instance (64-bit) and went through the installation process described here: https://github.com/mesos/mesos/wiki/.
When I tried creating a Mesos cluster as described in the Wiki page, I have no trouble creating the master. However, when I attempt to create the slave as described, I get the following:
I0605 22:36:28.665403 4030 logging.cpp:72] Logging to <stderr>
I0605 22:36:28.667832 4030 main.cpp:111] Creating "process" isolation module
I0605 22:36:28.667968 4030 main.cpp:119] Build: 2012-06-05 21:46:57 by ec2-user
I0605 22:36:28.667987 4030 main.cpp:120] Starting Mesos slave
I0605 22:36:28.668843 4045 slave.cpp:191] Slave started on 10.77.101.242:42842
I0605 22:36:28.668869 4045 slave.cpp:192] Slave resources: cpus=2; mem=6449
F0605 22:36:28.670444 4030 main.cpp:132] Check failed: detector.isSome() Failed
to create a master detector: Cannot parse '@0.0.0.0:0'
*** Check failure stack trace: ***
@ 0x7fdba0302ded google::LogMessage::Fail()
@ 0x7fdba03064a7 google::LogMessage::SendToLog()
@ 0x7fdba0305a26 google::LogMessage::Flush()
@ 0x7fdba03068fd google::LogMessageFatal::~LogMessageFatal()
@ 0x408150 main
@ 0x7fdb9ef30cdd __libc_start_main
@ 0x406229 (unknown)
./mesos-slave.sh: line 24: 4030 Aborted /tmp/mesos-0.9.0/build/
src/mesos-slave "${@}"
It appears the slave starts OK, but something fails right after that ...
Thanks,
Jim Donahue
Adobe
Re: Problem running Mesos tests
Posted by Benjamin Hindman <be...@eecs.berkeley.edu>.
To be clear, each master/slave right now has two ports, configurable via
--port and --webui_port.
Also, if you're running close to trunk you can play around with the new
javascript only webui that also uses these JSON endpoints. Point your
browser to ip:port (note 'port' not 'webui_port').
On Wed, Jun 6, 2012 at 9:21 AM, Benjamin Hindman <be...@eecs.berkeley.edu>wrote:
> Hi Jim,
>
> You can set the webui_port command line option. Try mesos-master --help.
>
> The webui gets all of its information via HTTP/JSON endpoints on the
> masters/slaves themselves, so you can definitely get access to the
> information. For the master do:
>
> $ curl 'ip:port/master/state.json'
>
> And if you're running off of trunk (or pretty close to it) you can do the
> following for the slave:
>
> $ curl 'ip:port/slave(1)/state.json'
>
> (Note that at these URLs might be changing slightly in the near future,
> but then be stable.)
>
> Hope that helps.
>
> Ben.
>
>
> On Wed, Jun 6, 2012 at 9:05 AM, Jim Donahue <jd...@adobe.com> wrote:
>
>> Thanks (and especially for the quick response)! I've got another
>> question:
>>
>> I'm going to be running Tomcat on the instances, so I'd like to use a
>> different port for the web UI that the masters provide -- is there a simple
>> way of changing the port? What I'd really like to do is to
>> programmatically get the information you display and collect it on a
>> dashboard (I'm building a service that will make it easy to set up and
>> manage multiple clusters). Is there a way to do that?
>>
>> Jim
>>
>>
>> -----Original Message-----
>> From: benjamin.hindman@gmail.com [mailto:benjamin.hindman@gmail.com] On
>> Behalf Of Benjamin Hindman
>> Sent: Tuesday, June 05, 2012 4:32 PM
>> To: mesos-dev@incubator.apache.org
>> Subject: Re: Problem running Mesos tests
>>
>> Hi Jim,
>>
>> My guess is that you used either '--master=mesos://master@ip:port'
>> or '--master=master@ip:port'. You should just use '--master=ip:port'.
>> I've
>> updated the github wiki. Sorry for any inconvenience.
>>
>> Ben.
>>
>>
>>
>>
>> On Tue, Jun 5, 2012 at 3:39 PM, Jim Donahue <jd...@adobe.com> wrote:
>>
>> > I've been trying to build an EC2 instance that has Mesos installed
>> (along
>> > with some other software). I started with the "standard" Amazon Linux
>> > instance (64-bit) and went through the installation process described
>> here:
>> > https://github.com/mesos/mesos/wiki/.
>> >
>> > When I tried creating a Mesos cluster as described in the Wiki page, I
>> > have no trouble creating the master. However, when I attempt to create
>> the
>> > slave as described, I get the following:
>> >
>> > I0605 22:36:28.665403 4030 logging.cpp:72] Logging to <stderr>
>> > I0605 22:36:28.667832 4030 main.cpp:111] Creating "process" isolation
>> > module
>> > I0605 22:36:28.667968 4030 main.cpp:119] Build: 2012-06-05 21:46:57 by
>> > ec2-user
>> >
>> > I0605 22:36:28.667987 4030 main.cpp:120] Starting Mesos slave
>> > I0605 22:36:28.668843 4045 slave.cpp:191] Slave started on
>> > 10.77.101.242:42842
>> > I0605 22:36:28.668869 4045 slave.cpp:192] Slave resources: cpus=2;
>> > mem=6449
>> > F0605 22:36:28.670444 4030 main.cpp:132] Check failed:
>> detector.isSome()
>> > Failed
>> > to create a master detector: Cannot parse '@0.0.0.0:0'
>> > *** Check failure stack trace: ***
>> > @ 0x7fdba0302ded google::LogMessage::Fail()
>> > @ 0x7fdba03064a7 google::LogMessage::SendToLog()
>> > @ 0x7fdba0305a26 google::LogMessage::Flush()
>> > @ 0x7fdba03068fd google::LogMessageFatal::~LogMessageFatal()
>> > @ 0x408150 main
>> > @ 0x7fdb9ef30cdd __libc_start_main
>> > @ 0x406229 (unknown)
>> > ./mesos-slave.sh: line 24: 4030 Aborted
>> > /tmp/mesos-0.9.0/build/
>> > src/mesos-slave "${@}"
>> >
>> > It appears the slave starts OK, but something fails right after that ...
>> >
>> > Thanks,
>> >
>> > Jim Donahue
>> > Adobe
>> >
>>
>
>
RE: Problem running Mesos tests
Posted by Jim Donahue <jd...@adobe.com>.
Hey, this is great! Thanks a lot!
Jim
-----Original Message-----
From: benjamin.hindman@gmail.com [mailto:benjamin.hindman@gmail.com] On Behalf Of Benjamin Hindman
Sent: Wednesday, June 06, 2012 9:21 AM
To: mesos-dev@incubator.apache.org
Subject: Re: Problem running Mesos tests
Hi Jim,
You can set the webui_port command line option. Try mesos-master --help.
The webui gets all of its information via HTTP/JSON endpoints on the
masters/slaves themselves, so you can definitely get access to the
information. For the master do:
$ curl 'ip:port/master/state.json'
And if you're running off of trunk (or pretty close to it) you can do the
following for the slave:
$ curl 'ip:port/slave(1)/state.json'
(Note that at these URLs might be changing slightly in the near future, but
then be stable.)
Hope that helps.
Ben.
On Wed, Jun 6, 2012 at 9:05 AM, Jim Donahue <jd...@adobe.com> wrote:
> Thanks (and especially for the quick response)! I've got another question:
>
> I'm going to be running Tomcat on the instances, so I'd like to use a
> different port for the web UI that the masters provide -- is there a simple
> way of changing the port? What I'd really like to do is to
> programmatically get the information you display and collect it on a
> dashboard (I'm building a service that will make it easy to set up and
> manage multiple clusters). Is there a way to do that?
>
> Jim
>
>
> -----Original Message-----
> From: benjamin.hindman@gmail.com [mailto:benjamin.hindman@gmail.com] On
> Behalf Of Benjamin Hindman
> Sent: Tuesday, June 05, 2012 4:32 PM
> To: mesos-dev@incubator.apache.org
> Subject: Re: Problem running Mesos tests
>
> Hi Jim,
>
> My guess is that you used either '--master=mesos://master@ip:port'
> or '--master=master@ip:port'. You should just use '--master=ip:port'. I've
> updated the github wiki. Sorry for any inconvenience.
>
> Ben.
>
>
>
>
> On Tue, Jun 5, 2012 at 3:39 PM, Jim Donahue <jd...@adobe.com> wrote:
>
> > I've been trying to build an EC2 instance that has Mesos installed (along
> > with some other software). I started with the "standard" Amazon Linux
> > instance (64-bit) and went through the installation process described
> here:
> > https://github.com/mesos/mesos/wiki/.
> >
> > When I tried creating a Mesos cluster as described in the Wiki page, I
> > have no trouble creating the master. However, when I attempt to create
> the
> > slave as described, I get the following:
> >
> > I0605 22:36:28.665403 4030 logging.cpp:72] Logging to <stderr>
> > I0605 22:36:28.667832 4030 main.cpp:111] Creating "process" isolation
> > module
> > I0605 22:36:28.667968 4030 main.cpp:119] Build: 2012-06-05 21:46:57 by
> > ec2-user
> >
> > I0605 22:36:28.667987 4030 main.cpp:120] Starting Mesos slave
> > I0605 22:36:28.668843 4045 slave.cpp:191] Slave started on
> > 10.77.101.242:42842
> > I0605 22:36:28.668869 4045 slave.cpp:192] Slave resources: cpus=2;
> > mem=6449
> > F0605 22:36:28.670444 4030 main.cpp:132] Check failed: detector.isSome()
> > Failed
> > to create a master detector: Cannot parse '@0.0.0.0:0'
> > *** Check failure stack trace: ***
> > @ 0x7fdba0302ded google::LogMessage::Fail()
> > @ 0x7fdba03064a7 google::LogMessage::SendToLog()
> > @ 0x7fdba0305a26 google::LogMessage::Flush()
> > @ 0x7fdba03068fd google::LogMessageFatal::~LogMessageFatal()
> > @ 0x408150 main
> > @ 0x7fdb9ef30cdd __libc_start_main
> > @ 0x406229 (unknown)
> > ./mesos-slave.sh: line 24: 4030 Aborted
> > /tmp/mesos-0.9.0/build/
> > src/mesos-slave "${@}"
> >
> > It appears the slave starts OK, but something fails right after that ...
> >
> > Thanks,
> >
> > Jim Donahue
> > Adobe
> >
>
Re: Problem running Mesos tests
Posted by Benjamin Hindman <be...@eecs.berkeley.edu>.
Hi Jim,
You can set the webui_port command line option. Try mesos-master --help.
The webui gets all of its information via HTTP/JSON endpoints on the
masters/slaves themselves, so you can definitely get access to the
information. For the master do:
$ curl 'ip:port/master/state.json'
And if you're running off of trunk (or pretty close to it) you can do the
following for the slave:
$ curl 'ip:port/slave(1)/state.json'
(Note that at these URLs might be changing slightly in the near future, but
then be stable.)
Hope that helps.
Ben.
On Wed, Jun 6, 2012 at 9:05 AM, Jim Donahue <jd...@adobe.com> wrote:
> Thanks (and especially for the quick response)! I've got another question:
>
> I'm going to be running Tomcat on the instances, so I'd like to use a
> different port for the web UI that the masters provide -- is there a simple
> way of changing the port? What I'd really like to do is to
> programmatically get the information you display and collect it on a
> dashboard (I'm building a service that will make it easy to set up and
> manage multiple clusters). Is there a way to do that?
>
> Jim
>
>
> -----Original Message-----
> From: benjamin.hindman@gmail.com [mailto:benjamin.hindman@gmail.com] On
> Behalf Of Benjamin Hindman
> Sent: Tuesday, June 05, 2012 4:32 PM
> To: mesos-dev@incubator.apache.org
> Subject: Re: Problem running Mesos tests
>
> Hi Jim,
>
> My guess is that you used either '--master=mesos://master@ip:port'
> or '--master=master@ip:port'. You should just use '--master=ip:port'. I've
> updated the github wiki. Sorry for any inconvenience.
>
> Ben.
>
>
>
>
> On Tue, Jun 5, 2012 at 3:39 PM, Jim Donahue <jd...@adobe.com> wrote:
>
> > I've been trying to build an EC2 instance that has Mesos installed (along
> > with some other software). I started with the "standard" Amazon Linux
> > instance (64-bit) and went through the installation process described
> here:
> > https://github.com/mesos/mesos/wiki/.
> >
> > When I tried creating a Mesos cluster as described in the Wiki page, I
> > have no trouble creating the master. However, when I attempt to create
> the
> > slave as described, I get the following:
> >
> > I0605 22:36:28.665403 4030 logging.cpp:72] Logging to <stderr>
> > I0605 22:36:28.667832 4030 main.cpp:111] Creating "process" isolation
> > module
> > I0605 22:36:28.667968 4030 main.cpp:119] Build: 2012-06-05 21:46:57 by
> > ec2-user
> >
> > I0605 22:36:28.667987 4030 main.cpp:120] Starting Mesos slave
> > I0605 22:36:28.668843 4045 slave.cpp:191] Slave started on
> > 10.77.101.242:42842
> > I0605 22:36:28.668869 4045 slave.cpp:192] Slave resources: cpus=2;
> > mem=6449
> > F0605 22:36:28.670444 4030 main.cpp:132] Check failed: detector.isSome()
> > Failed
> > to create a master detector: Cannot parse '@0.0.0.0:0'
> > *** Check failure stack trace: ***
> > @ 0x7fdba0302ded google::LogMessage::Fail()
> > @ 0x7fdba03064a7 google::LogMessage::SendToLog()
> > @ 0x7fdba0305a26 google::LogMessage::Flush()
> > @ 0x7fdba03068fd google::LogMessageFatal::~LogMessageFatal()
> > @ 0x408150 main
> > @ 0x7fdb9ef30cdd __libc_start_main
> > @ 0x406229 (unknown)
> > ./mesos-slave.sh: line 24: 4030 Aborted
> > /tmp/mesos-0.9.0/build/
> > src/mesos-slave "${@}"
> >
> > It appears the slave starts OK, but something fails right after that ...
> >
> > Thanks,
> >
> > Jim Donahue
> > Adobe
> >
>
RE: Problem running Mesos tests
Posted by Jim Donahue <jd...@adobe.com>.
Thanks (and especially for the quick response)! I've got another question:
I'm going to be running Tomcat on the instances, so I'd like to use a different port for the web UI that the masters provide -- is there a simple way of changing the port? What I'd really like to do is to programmatically get the information you display and collect it on a dashboard (I'm building a service that will make it easy to set up and manage multiple clusters). Is there a way to do that?
Jim
-----Original Message-----
From: benjamin.hindman@gmail.com [mailto:benjamin.hindman@gmail.com] On Behalf Of Benjamin Hindman
Sent: Tuesday, June 05, 2012 4:32 PM
To: mesos-dev@incubator.apache.org
Subject: Re: Problem running Mesos tests
Hi Jim,
My guess is that you used either '--master=mesos://master@ip:port'
or '--master=master@ip:port'. You should just use '--master=ip:port'. I've
updated the github wiki. Sorry for any inconvenience.
Ben.
On Tue, Jun 5, 2012 at 3:39 PM, Jim Donahue <jd...@adobe.com> wrote:
> I've been trying to build an EC2 instance that has Mesos installed (along
> with some other software). I started with the "standard" Amazon Linux
> instance (64-bit) and went through the installation process described here:
> https://github.com/mesos/mesos/wiki/.
>
> When I tried creating a Mesos cluster as described in the Wiki page, I
> have no trouble creating the master. However, when I attempt to create the
> slave as described, I get the following:
>
> I0605 22:36:28.665403 4030 logging.cpp:72] Logging to <stderr>
> I0605 22:36:28.667832 4030 main.cpp:111] Creating "process" isolation
> module
> I0605 22:36:28.667968 4030 main.cpp:119] Build: 2012-06-05 21:46:57 by
> ec2-user
>
> I0605 22:36:28.667987 4030 main.cpp:120] Starting Mesos slave
> I0605 22:36:28.668843 4045 slave.cpp:191] Slave started on
> 10.77.101.242:42842
> I0605 22:36:28.668869 4045 slave.cpp:192] Slave resources: cpus=2;
> mem=6449
> F0605 22:36:28.670444 4030 main.cpp:132] Check failed: detector.isSome()
> Failed
> to create a master detector: Cannot parse '@0.0.0.0:0'
> *** Check failure stack trace: ***
> @ 0x7fdba0302ded google::LogMessage::Fail()
> @ 0x7fdba03064a7 google::LogMessage::SendToLog()
> @ 0x7fdba0305a26 google::LogMessage::Flush()
> @ 0x7fdba03068fd google::LogMessageFatal::~LogMessageFatal()
> @ 0x408150 main
> @ 0x7fdb9ef30cdd __libc_start_main
> @ 0x406229 (unknown)
> ./mesos-slave.sh: line 24: 4030 Aborted
> /tmp/mesos-0.9.0/build/
> src/mesos-slave "${@}"
>
> It appears the slave starts OK, but something fails right after that ...
>
> Thanks,
>
> Jim Donahue
> Adobe
>
Re: Problem running Mesos tests
Posted by Benjamin Hindman <be...@eecs.berkeley.edu>.
Hi Jim,
My guess is that you used either '--master=mesos://master@ip:port'
or '--master=master@ip:port'. You should just use '--master=ip:port'. I've
updated the github wiki. Sorry for any inconvenience.
Ben.
On Tue, Jun 5, 2012 at 3:39 PM, Jim Donahue <jd...@adobe.com> wrote:
> I've been trying to build an EC2 instance that has Mesos installed (along
> with some other software). I started with the "standard" Amazon Linux
> instance (64-bit) and went through the installation process described here:
> https://github.com/mesos/mesos/wiki/.
>
> When I tried creating a Mesos cluster as described in the Wiki page, I
> have no trouble creating the master. However, when I attempt to create the
> slave as described, I get the following:
>
> I0605 22:36:28.665403 4030 logging.cpp:72] Logging to <stderr>
> I0605 22:36:28.667832 4030 main.cpp:111] Creating "process" isolation
> module
> I0605 22:36:28.667968 4030 main.cpp:119] Build: 2012-06-05 21:46:57 by
> ec2-user
>
> I0605 22:36:28.667987 4030 main.cpp:120] Starting Mesos slave
> I0605 22:36:28.668843 4045 slave.cpp:191] Slave started on
> 10.77.101.242:42842
> I0605 22:36:28.668869 4045 slave.cpp:192] Slave resources: cpus=2;
> mem=6449
> F0605 22:36:28.670444 4030 main.cpp:132] Check failed: detector.isSome()
> Failed
> to create a master detector: Cannot parse '@0.0.0.0:0'
> *** Check failure stack trace: ***
> @ 0x7fdba0302ded google::LogMessage::Fail()
> @ 0x7fdba03064a7 google::LogMessage::SendToLog()
> @ 0x7fdba0305a26 google::LogMessage::Flush()
> @ 0x7fdba03068fd google::LogMessageFatal::~LogMessageFatal()
> @ 0x408150 main
> @ 0x7fdb9ef30cdd __libc_start_main
> @ 0x406229 (unknown)
> ./mesos-slave.sh: line 24: 4030 Aborted
> /tmp/mesos-0.9.0/build/
> src/mesos-slave "${@}"
>
> It appears the slave starts OK, but something fails right after that ...
>
> Thanks,
>
> Jim Donahue
> Adobe
>