Unable to run latest windows mesas-agent

[Rinaldo Digiorgio]

Hi,

I built a new windows mesas-agent.exe a few minutes ago and I am getting the 
following error. I don’t think it is a port issue.

I0913 18:07:22.359485 12992 slave.cpp:1692] Launching task 
'windows-hello-world.83d04669-7a17-11e6-ae4e-0021f6964572' for framework 
03b7ca4b-1a97-4044-ba65-9a92458987ad-0026
W0913 18:07:22.359485 12992 slave.cpp:1760] Ignoring running task 
'windows-hello-world.83d04669-7a17-11e6-ae4e-0021f6964572' of framework 
03b7ca4b-1a97-4044-ba65-9a92458987ad-0026 because the framework is terminating
I0913 18:07:22.359485 12992 slave.cpp:4660] Cleaning up framework 
03b7ca4b-1a97-4044-ba65-9a92458987ad-0026
I0913 18:07:22.359485  3600 status_update_manager.cpp:285] Closing status 
update streams for framework 03b7ca4b-1a97-4044-ba65-9a92458987ad-0026
E0913 18:07:22.359485 12992 slave.cpp:5350] Failed to find the mtime of 
'C:\mesos\slaves\1eb74be0-b996-4da5-a3a7-ee8ea7728dc5-S20\frameworks\03b7ca4b-1a97-4044-ba65-9a92458987ad-0026':
 Error invoking stat for 
'C:\mesos\slaves\1eb74be0-b996-4da5-a3a7-ee8ea7728dc5-S20\frameworks\03b7ca4b-1a97-4044-ba65-9a92458987ad-0026':
 No such file or directory
I0913 18:07:22.359485  8988 gc.cpp:55] Scheduling 
'C:\mesos\meta\slaves\1eb74be0-b996-4da5-a3a7-ee8ea7728dc5-S20\frameworks\03b7ca4b-1a97-4044-ba65-9a92458987ad-0026'
 for gc 6.9583928296days in the future
I0913 18:07:22.359485 12992 slave.cpp:783] Agent terminating
ABORT: 
(C:\jenkins\workspace\mesos-windows-build\3rdparty\stout\include\stout/os/windows/socket.hpp:136):
 Not expecting 'getsockopt' to fail when passed a valid socket

Windows Build on Jenkins almost working

[Rinaldo Digiorgio]

Hi,

 The build fails with the following.

 http_parser.lib(http_parser.obj) : warning LNK4217: locally defined 
symbol memchr imported in function http_parser_execute [C:\Program Files 
(x86)\Jenkins\workspace\mesos-agent-windows\build\3rdparty\libprocess\src\tests\process_tests.vcxproj]


   "C:\Program Files 
(x86)\Jenkins\workspace\mesos-agent-windows\build\Mesos.sln" (stout_tests;Build 
target) (1) ->
   "C:\Program Files 
(x86)\Jenkins\workspace\mesos-agent-windows\build\src\mesos-1.0.0.vcxproj.metaproj"
 (default target) (28) ->
   "C:\Program Files 
(x86)\Jenkins\workspace\mesos-agent-windows\build\3rdparty\zookeeper-06d3f3f.vcxproj.metaproj"
 (default target) (31) ->
   "C:\Program Files 
(x86)\Jenkins\workspace\mesos-agent-windows\build\3rdparty\zookeeper-06d3f3f.vcxproj"
 (default target) (40) ->
   (CustomBuild target) -> 
 C:\Program Files 
(x86)\MSBuild\Microsoft.Cpp\v4.0\V140\Microsoft.CppCommon.targets(171,5): error 
MSB6006: "cmd.exe" exited with code 9009. [C:\Program Files 
(x86)\Jenkins\workspace\mesos-agent-windows\build\3rdparty\zookeeper-06d3f3f.vcxproj]

1474 Warning(s)
1 Error(s)

Time Elapsed 00:24:27.44


  I see this.

CMake Warning at CMakeLists.txt:52 (message):
  Both `ENABLE_LIBEVENT` and `REBUNDLED` (set to TRUE by default) flags have
  been set.  But, libevent does not come rebundled in Mesos, so it must be
  downloaded.


CMake Warning at CMakeLists.txt:61 (message):
  The current supported version of ZK does not compile on Windows, and does
  not come rebundled in the Mesos repository.  It must be downloaded from the
  Internet, even though the `REBUNDLED` flag was set.


If I need to install libevent, what version and where is a good place for it?

Rinaldo

Re: OS X latency issue when run as a plist

[Rinaldo Digiorgio]

> On Jul 13, 2016, at 9:20 PM, Rodrick Brown <rodr...@orchard-app.com> wrote:
> 
> Have you tried using something like supervisord? Or the slew of other process 
> launchers available for *nix. 
> 
Thanks I had no idea that there were alternatives to launchd. Thanks will look 
into it and report back for the next person.
> Check brew.
> 
> I would look to that as an interim solution if the plist method remains 
> problematic. 
> 
> Get Outlook for iOS <https://aka.ms/o0ukef>
> 
> 
> 
> On Wed, Jul 13, 2016 at 7:44 AM -0400, "Rinaldo Digiorgio" 
> <rdigior...@pace.edu <mailto:rdigior...@pace.edu>> wrote:
> 
> Hi,
> 
>   There have been prior discussions on the list about the OS X Latency 
> issue. I had filed a bug here:
> 
>   https://issues.apache.org/jira/browse/MESOS-5589 
> <https://issues.apache.org/jira/browse/MESOS-5589>
> 
>   We have found that the root cause is starting the mesos application in 
> the background using a plist entry.  If you launch the mesos agent from a 
> terminal it works fine.  We have tried to get a plist (not an app) to work 
> and none of the documented settings in launchd remove the latency issue.
> 
>   
> https://developer.apple.com/legacy/library/documentation/Darwin/Reference/ManPages/man5/launchd.plist.5.html
>  
> <https://developer.apple.com/legacy/library/documentation/Darwin/Reference/ManPages/man5/launchd.plist.5.html>
> 
>   The settings we tried are:
> 
>ProcessType 
>  This optional key describes, at a high level, the intended purpose of 
> the job.  The system will apply
>  resource limits based on what kind of job it is. If left unspecified, 
> the system will apply light
>  resource limits to the job, throttling its CPU usage and I/O bandwidth. 
> The following are valid values:
> 
>Background
>Background jobs are generally processes that do work that was not 
> directly requested by the user.
>The resource limits applied to Background jobs are intended to 
> prevent them from disrupting the
>user experience.
> 
>Standard
>Standard jobs are equivalent to no ProcessType being set.
> 
>Adaptive
>Adaptive jobs move between the Background and Interactive 
> classifications based on activity over
>XPC connections. See xpc_transaction_begin(3) 
> <https://developer.apple.com/legacy/library/documentation/Darwin/Reference/ManPages/man3/xpc_transaction_begin.3.html#//apple_ref/doc/man/3/xpc_transaction_begin>
>  for details.
> 
>Interactive
>Interactive jobs run with the same resource limitations as apps, 
> that is to say, none. Interac-tive Interactive
>tive jobs are critical to maintaining a responsive user 
> experience, and this key should only be
>used if an app's ability to be responsive depends on it, and 
> cannot be made Adaptive.
> 
> 
> The mesos agent works correctly if you start it as a GUI app. This leaves an 
> icon on the screen. One can live with it but it is an indication of the lack 
> of proper documentation from apple and or utter lack of understanding of 
> background application on the Desktop OS known as OS X.  If someone has a 
> plist solution please share it. It is not reasonable to start mesos agents 
> from a terminal session or cron, the operating system should manage startup 
> and shutdown.
> 
> Rinaldo
> 
> 
> 
>   
> 
> NOTICE TO RECIPIENTS: This communication is confidential and intended for the 
> use of the addressee only. If you are not an intended recipient of this 
> communication, please delete it immediately and notify the sender by return 
> email. Unauthorized reading, dissemination, distribution or copying of this 
> communication is prohibited. This communication does not constitute an offer 
> to sell or a solicitation of an indication of interest to purchase any loan, 
> security or any other financial product or instrument, nor is it an offer to 
> sell or a solicitation of an indication of interest to purchase any products 
> or services to any persons who are prohibited from receiving such information 
> under applicable law. The contents of this communication may not be accurate 
> or complete and are subject to change without notice. As such, Orchard App, 
> Inc. (including its subsidiaries and affiliates, "Orchard") makes no 
> representation regarding the accuracy or completeness of the information 
> contained herein. The intended recipient is advised to consult its own 
> professional advisors, including those specializing in legal, tax and 
> accounting matters. Orchard does not provide legal, tax or accounting advice.

OS X latency issue when run as a plist

[Rinaldo Digiorgio]

Hi,

There have been prior discussions on the list about the OS X Latency 
issue. I had filed a bug here:

https://issues.apache.org/jira/browse/MESOS-5589 


We have found that the root cause is starting the mesos application in 
the background using a plist entry.  If you launch the mesos agent from a 
terminal it works fine.  We have tried to get a plist (not an app) to work and 
none of the documented settings in launchd remove the latency issue.


https://developer.apple.com/legacy/library/documentation/Darwin/Reference/ManPages/man5/launchd.plist.5.html
 


The settings we tried are:

 ProcessType 
 This optional key describes, at a high level, the intended purpose of the 
job.  The system will apply
 resource limits based on what kind of job it is. If left unspecified, the 
system will apply light
 resource limits to the job, throttling its CPU usage and I/O bandwidth. 
The following are valid values:

   Background
   Background jobs are generally processes that do work that was not 
directly requested by the user.
   The resource limits applied to Background jobs are intended to 
prevent them from disrupting the
   user experience.

   Standard
   Standard jobs are equivalent to no ProcessType being set.

   Adaptive
   Adaptive jobs move between the Background and Interactive 
classifications based on activity over
   XPC connections. See xpc_transaction_begin(3) 

 for details.

   Interactive
   Interactive jobs run with the same resource limitations as apps, 
that is to say, none. Interac-tive Interactive
   tive jobs are critical to maintaining a responsive user experience, 
and this key should only be
   used if an app's ability to be responsive depends on it, and cannot 
be made Adaptive.


The mesos agent works correctly if you start it as a GUI app. This leaves an 
icon on the screen. One can live with it but it is an indication of the lack of 
proper documentation from apple and or utter lack of understanding of 
background application on the Desktop OS known as OS X.  If someone has a plist 
solution please share it. It is not reasonable to start mesos agents from a 
terminal session or cron, the operating system should manage startup and 
shutdown.

Rinaldo

Windows Build

[Rinaldo Digiorgio]

Hi,
Would someone be able to suggest how to get started with building mesos 
on windows. I am under the assumption that the windows branch is not integrated 
into the current 1.* RC.  

Rinaldo

Re: OSX 10.10.5 and mesos 0.28.1 -- 10 to 20 X difference in sleep() method compared to non mesos

[Rinaldo Digiorgio]

I performed the following tests

Built the latest tip 1.0 on OS X 10.11.4,  2.3 GHz Intel Core i7 with 
16GB and load average of 1+ I ran the test program, SleepLatency.java  with 
Java 1.8 and 1.7 and I am getting very poor results. I am unable to reproduce 
the results reported by haosdent .  I have also increased the priority of the 
mesos agent and mesos master.  Is there something I am missing here in the 
config.

I have the master and the agent running on the same machine.

Not using zookeeper

Using localhost

No special options for the master and the agent

0 894 1 0 2:15PM ?? 0:29.22 /usr/local/sbin/mesos-master 
--work_dir=/scratch/mesos/master --ip=127.0.0.1 
0 896 1 0 2:15PM ?? 1:15.55 /usr/local/sbin/mesos-slave 
--master=127.0.0.1:5050 --work_dir=/scratch/mesos/slave

Turned off power nap


Rinaldo


This is the application


 public class SleepLatency {
static final int COUNT = 100;
static final long DELAY = 10L;

public static void main(String[] args) throws Exception {
long tstart = System.currentTimeMillis();
for (int i = 0; i < COUNT; i++) {
long t0 = System.currentTimeMillis();
Thread.sleep(DELAY);
long t1 = System.currentTimeMillis();
System.out.printf("loop %3d delay %4d ms%n", i, t1 - t0);
}
long tfinish = System.currentTimeMillis();
System.out.printf("total time = %5d ms%n", tfinish - tstart);
}
 }

and how it was launched

mesos-execute --master="127.0.0.1:5050" --name="SleepLatency" --command="cd 
/tmp && java SleepLatency”


Rinaldo

> On Jun 6, 2016, at 11:28 AM, haosdent  wrote:
> 
> yes, mine is a laptop.
> 
> On Mon, Jun 6, 2016 at 11:23 PM, DiGiorgio, Mr. Rinaldo S. 
>  wrote:
> 
>> On Jun 6, 2016, at 11:08, haosdent  wrote:
>> 
>> I use OS X 10.11.4 but I think it should not related to this minor different 
>> on OS X version.
>> For settings, I disable `Enable Power Nap while plugged into a power 
>> adapter` in `Energy Saver`.
> 
> I assume you have a laptop, I have retested with power nap being set to off 
> on an iMac and it did not make a difference. I will try it on a laptop.  The 
> other 15 Macs are all towers. Maybe it is a clue, works on laptops but not on 
> desktops.
>> On Mon, Jun 6, 2016 at 10:57 PM, DiGiorgio, Mr. Rinaldo S. 
>>  wrote:
> 
> 
>> Thanks James and Haosdent,
>> 
>> I built my own version of mesos 1.0 and installed it on 10.11.5 and I am 
>> getting the same results as 10.10.5 with 0.28.1
>> 
>> Do either of you remember what you may have set on your OSX machines to 
>> default some of the desktop settings that may be causing this issue?
>> 
>> 
>> We are already turning off many desktop features since they are not relevant 
>> for our use case.
>> 
>> 
>> 
>> Received SUBSCRIBED event
>> Subscribed executor on dhcp-adc-twvpn-3-vpnpool-10-154-101-79.vpn.oracle.com
>> Received LAUNCH event
>> Starting task test-sleep
>> sh -c 'cd /tmp && java SleepLatency'
>> Forked command at 2907
>> loop   0 delay   35 ms
>> loop   1 delay  108 ms
>> loop   2 delay  104 ms
>> loop   3 delay   57 ms
>> loop   4 delay  104 ms
>> loop   5 delay   93 ms
>> loop   6 delay   12 ms
>> loop   7 delay   17 ms
>> loop   8 delay  105 ms
>> loop   9 delay  109 ms
>> 
>> 
>> > On Jun 4, 2016, at 05:07, haosdent  wrote:
>> >
>> > Hi, Rinaldo. I test your problem in my local Mesos (run in my mac). It 
>> > looks normal in my side. I started it by
>> >
>> > ```
>> > mesos-execute --master="localhost:5050" --name="test-sleep" --command="cd 
>> > /tmp && java SleepLatency"
>> > ```
>> >
>> > ```
>> > Registered executor on localhost
>> > Starting task test-sleep
>> > sh -c 'cd /tmp && java SleepLatency'
>> > Forked command at 42480
>> > loop   0 delay   11 ms
>> > loop   1 delay   12 ms
>> > loop   2 delay   11 ms
>> > loop   3 delay   13 ms
>> > loop   4 delay   12 ms
>> > loop   5 delay   12 ms
>> > loop   6 delay   12 ms
>> > loop   7 delay   12 ms
>> > loop   8 delay   12 ms
>> > loop   9 delay   11 ms
>> > loop  10 delay   12 ms
>> > loop  11 delay   13 ms
>> > loop  12 delay   11 ms
>> > loop  13 delay   11 ms
>> > loop  14 delay   12 ms
>> > loop  15 delay   12 ms
>> > loop  16 delay   13 ms
>> > loop  17 delay   12 ms
>> > loop  18 delay   11 ms
>> > loop  19 delay   12 ms
>> > loop  20 delay   12 ms
>> > loop  21 delay   11 ms
>> > loop  22 delay   12 ms
>> > loop  23 delay   11 ms
>> > loop  24 delay   12 ms
>> > loop  25 delay   12 ms
>> > loop  26 delay   12 ms
>> > loop  27 delay   12 ms
>> > loop  28 delay   13 ms
>> > loop  29 delay   12 ms
>> > loop  30 delay   12 ms
>> > loop  31 delay   11 ms
>> > loop  32 delay   11 ms
>> > loop  33 delay   11 ms
>> > loop  34 delay   13 ms
>> > loop  35 delay   10 ms
>> > loop  36 delay   12 ms
>> > loop  37 delay   11 ms
>> > loop  38 delay   

Is there a one minute timeout on URI downloads or extractions?

[Rinaldo Digiorgio]

Hi,

I have several URI files configured in Marathon for download to the 
agent.  A mesos task starts and terminates consistently when trying to download 
the files.  I see the following message in the log.


I0503 20:24:38.330714  2756 slave.cpp:4322] Terminating executor 
‘'test.97c0bc52-11a7-11e6-99c8-56847afe9799' of framework 
e093b9e5-ad72-40e0-b686-9d41691141e5-1265' because it did not register within 
1mins

stderr shows some files being unzipped


Rinaldo

Re: How to add other file systems to an agent

[Rinaldo Digiorgio]

> On May 3, 2016, at 7:31 AM, Dick Davies <d...@hellooperator.net> wrote:
> 
> I'd imagine it's reporting whatever partition the --work-dir argument
> on the slave is set to (sandboxes live under that directory).
Thanks


I see that there is support for attaching storage:  
http://mesos.apache.org/documentation/latest/persistent-volume/ 
<http://mesos.apache.org/documentation/latest/persistent-volume/>





> On 3 May 2016 at 12:21, Rinaldo Digiorgio <rdigior...@pace.edu> wrote:
>> Hi,
>> 
>>I have a configuration with a root file system and other file 
>> systems. When I start an agent, the agent only reports the disk on the root 
>> file system.  Is there a way to specify a list of file systems to include as 
>> resources of the agent when it starts? I checked the agent options.
>> 
>> 
>> Rinaldo

How to add other file systems to an agent

[Rinaldo Digiorgio]

Hi,

I have a configuration with a root file system and other file systems. 
When I start an agent, the agent only reports the disk on the root file system. 
 Is there a way to specify a list of file systems to include as resources of 
the agent when it starts? I checked the agent options.


Rinaldo

Unable to build 2.6 on OS X

[Rinaldo Digiorgio]

I am trying to build 0.26 on OS/X  10.11.2 and failing in the configure step.
I think all the required libraries are installed

configure: error: cannot find libapr-1 headers
---
libapr-1 is required for mesos to build.
—


../configure  –-with-apr=/usr/local/Cellar/apr/1.5.2/libexec/
configure: error: invalid variable name: `–-with-apr’


configure —help shows the following so the option should be accepted

  --with-apr=[=DIR]   specify where to locate the apr-1 library


Rinaldo

Re: Unable to build 2.6 on OS X

[Rinaldo Digiorgio]

> On Jan 29, 2016, at 12:30 AM, Shuai Lin  wrote:
> 
> A googling of "configure: error: invalid variable name" leads me to 
> http://askubuntu.com/a/590679  . The reason: 
> the first dash in your '--with-apr' is not correct typed.

Thanks that was it. I did a copy and paste I guess from a page that had a non 
printable. I checked the mesos page and it doesn’t have this issue.