Message boards : General Discussion : The BOINC VM application
Message board moderation

To post messages, you must log in.

1 · 2 · 3 · Next

AuthorMessage
Profile Laurence
Project administrator
Project developer
Project tester
Avatar

Send message
Joined: 12 Sep 14
Posts: 1064
Credit: 328,405
RAC: 184
Message 6447 - Posted: 19 Jul 2019, 11:46:23 UTC

The BOINC VM application aims to provide a generic platform for running Linux applications.

The goal is to make life simple for everyone; the volunteer, the administrator and the scientist. Linux containers are good for the scientist as the applications they build in a containers should be able to run anywhere. Our recent experiences with native applications (both ATLAS and Theory) as well as boinc2docker apps show that containers work. The limitation is that this will only work for Linux or a VM is required that may be both troublesome and inefficient. The BOINC VM application addresses these issues by running the BOINC client in the VM. This results in two instances of the client being used. The one on the host focuses on the resource sharing policy and the one in the guest focuses on task scheduling.

The initial version of the BOINC VM application only connects to LHC@home but once pull requests 3191 3209 and 3214 have been added, it should be possible to manage the BOINC client in the Guest via a Web browser.

We look forward to your feedback.
ID: 6447 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
maeax

Send message
Joined: 22 Apr 16
Posts: 664
Credit: 1,805,961
RAC: 2,479
Message 6448 - Posted: 19 Jul 2019, 12:44:03 UTC
Last modified: 19 Jul 2019, 12:56:57 UTC

First task in Win10 pro(x64)
In RDP under F1:
========= Tasks =========== shown in many lines (every few minutes one line).
F2, F3 or F4 are empty.
https://lhcathomedev.cern.ch/lhcathome-dev/results.php?hostid=1377
Now is a Host -localhost generated:
https://lhcathomedev.cern.ch/lhcathome-dev/show_host_detail.php?hostid=3856
ID: 6448 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer
Project tester
Avatar

Send message
Joined: 12 Sep 14
Posts: 1064
Credit: 328,405
RAC: 184
Message 6449 - Posted: 19 Jul 2019, 13:24:21 UTC - in response to Message 6448.  

It should pick up a task after about four or five loops. There was an issue with not enough memory with v0.01. I have now discovered there may be an issue due to the lack of swap.
ID: 6449 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer
Project tester
Avatar

Send message
Joined: 12 Sep 14
Posts: 1064
Credit: 328,405
RAC: 184
Message 6450 - Posted: 19 Jul 2019, 14:24:34 UTC - in response to Message 6449.  

One issue was the memory set in the plan class. I have also add some swap. I am not getting tasks now so will need to investigate. You can access the VM as your authenticator has been set as the root password.
ID: 6450 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer
Project tester
Avatar

Send message
Joined: 12 Sep 14
Posts: 1064
Credit: 328,405
RAC: 184
Message 6451 - Posted: 19 Jul 2019, 14:45:42 UTC - in response to Message 6450.  

It is now working for me. The problem was due to the memory definition in the plan_classes.
ID: 6451 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
maeax

Send message
Joined: 22 Apr 16
Posts: 664
Credit: 1,805,961
RAC: 2,479
Message 6452 - Posted: 19 Jul 2019, 15:07:11 UTC

Have purged 0.02 and resetted the Project and got now a 0.03 Task.
Shows this ==== Tasks === lines again.
Will waiting a few hours.
ID: 6452 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Ray Murray
Avatar

Send message
Joined: 13 Apr 15
Posts: 138
Credit: 2,945,852
RAC: 0
Message 6453 - Posted: 19 Jul 2019, 19:13:57 UTC
Last modified: 19 Jul 2019, 20:03:58 UTC

version 0.03
Started off with 4 =====Tasks==== lines a per Maeax but after a few minutes it looked like this screengrab that I took 20mins in



Numbers in top and bottom sections are similar but I have set to use only 1 CPU so I'm not sure if it is trying to run 2 jobs on that 1 core?
It's estimating 4days+ to completion but that estimate is only ticking down 1 second for every 2 seconds real time.

Looking at my Tasks, I see one Task running on my host and another on a new localhost, presumably within the VM created by Boinc on the real host. Will that result in a new computer id being generated each time Boinc starts a new task and therefore a new VM or will the localhost id pass to the new VM?
ID: 6453 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Laurence
Project administrator
Project developer
Project tester
Avatar

Send message
Joined: 12 Sep 14
Posts: 1064
Credit: 328,405
RAC: 184
Message 6456 - Posted: 19 Jul 2019, 21:47:03 UTC - in response to Message 6453.  

Great! Let's see how it goes. Looks like I am returning valids.
ID: 6456 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
maeax

Send message
Joined: 22 Apr 16
Posts: 664
Credit: 1,805,961
RAC: 2,479
Message 6457 - Posted: 20 Jul 2019, 4:01:49 UTC

Ray,
have Virtualbox 5.2.26. Seem so, that there are no tasks running up to now. 12 hours left.
Seeing also a separate localhost
ID: 6457 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Ray Murray
Avatar

Send message
Joined: 13 Apr 15
Posts: 138
Credit: 2,945,852
RAC: 0
Message 6458 - Posted: 20 Jul 2019, 7:28:52 UTC
Last modified: 20 Jul 2019, 8:18:58 UTC

Morning update
My shiny new localhost has successfully completed 5 jobs overnight, all credited and showing in McPlots 8¬) and 3 tasks in progress.
Console window shows Task 1) complete and Tasks 2) and 3) running
all still within the original Boinc Task on the real host. Still not sure why 2 tasks appear to be active when I specified only 1 core and indeed there is only 1 core in the VBox panel.
Still 4 days remaining after 13hrs runtime.
Seems to be running well enough to allow my other host to try one.
Also looks good straight away, or at least it looks the same. Maybe too early to judge outright success until we see what happens when the Boinc portion of the Task ends.

Win 10
Boinc 7.14.2
VBox 6.0.10
ID: 6458 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 8 Apr 15
Posts: 751
Credit: 11,608,083
RAC: 1,481
Message 6459 - Posted: 20 Jul 2019, 10:37:29 UTC

I have had 3 of these tasks running 2-core each for the last 8 hours (3:30am now) so I will see how they look when I wake up later.......when I checked my list of computers here as I always do I saw that *localhost* and almost............well I guess I should have checked to see if there was a thread for this first
ID: 6459 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Crystal Pellet
Volunteer tester

Send message
Joined: 13 Feb 15
Posts: 1180
Credit: 815,336
RAC: 238
Message 6460 - Posted: 20 Jul 2019, 13:10:28 UTC
Last modified: 20 Jul 2019, 13:46:32 UTC

My first task stays in downloading state. Not any CPU used within the VM.


I closed that task and requested a new one. Same result. Is this maybe cause there are no sub-jobs?
ID: 6460 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Crystal Pellet
Volunteer tester

Send message
Joined: 13 Feb 15
Posts: 1180
Credit: 815,336
RAC: 238
Message 6461 - Posted: 20 Jul 2019, 16:33:27 UTC

The files needed

wrapper_2019_03_02_x86_64-linux
cranky-0.1.1
cvmfs-mini-0.1-amd64.tgz
Theory_job_2018_12_12.xml
2279-795411-80.run


are not downloaded into the project directory, although the boinc process is running within the VM.
ID: 6461 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
maeax

Send message
Joined: 22 Apr 16
Posts: 664
Credit: 1,805,961
RAC: 2,479
Message 6462 - Posted: 20 Jul 2019, 16:57:45 UTC - in response to Message 6458.  

Ray,
your localhost is a SL Linux with running Theory -native V. 4.35 and not Boinc VM App, sorry.
ID: 6462 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Ray Murray
Avatar

Send message
Joined: 13 Apr 15
Posts: 138
Credit: 2,945,852
RAC: 0
Message 6463 - Posted: 20 Jul 2019, 18:33:18 UTC - in response to Message 6462.  

Hi Maeax,
My understanding of Laurence's original post is that the Boinc VM App creates the localhost SL Linux virtually (just as it has created this host on your own machine, although you appear not to have gotten any jobs there for some reason) in order to allow running of Native (Linux) tasks, which I can't normally run on my Windows hosts.

I believe the Boinc instance within the VM does all the work while the Boinc outside, on the physical host does nothing other than create the VM and perhaps destroy it on completion.
I'm open to correction or clarification, but that's my interpretation of what's going on.

My 2 localhost Linux hosts on different Windows machines have successfully completed 8 and 6 tasks respectively. There might now be more confusion as these Virtual hosts are running Tasks which would have been called Jobs had they been run within a Boinc Task 8\
ID: 6463 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
maeax

Send message
Joined: 22 Apr 16
Posts: 664
Credit: 1,805,961
RAC: 2,479
Message 6464 - Posted: 20 Jul 2019, 18:50:44 UTC
Last modified: 20 Jul 2019, 18:58:04 UTC

Yes Ray,
there is a problem of understanding how this Boinc VM App is working.
Have searching the stderr.txt in the Boinc Folder of my running VM App and saw this lines:
2019-07-19 17:04:01 (1356): Guest Log: [INFO] Waiting for the BOINC client to start ...

2019-07-19 17:04:08 (1356): Guest Log: [INFO] Attaching to https://lhcathomedev.cern.ch/lhcathome-dev ...

2019-07-19 18:44:37 (1356): Status Report: Elapsed Time: '6000.000000'
2019-07-19 18:44:37 (1356): Status Report: CPU Time: '100.406250'
2019-07-19 20:26:14 (1356): Status Report: Elapsed Time: '12000.000000'
2019-07-19 20:26:14 (1356): Status Report: CPU Time: '188.671875'
2019-07-19 22:11:38 (1356): Status Report: Elapsed Time: '18000.088625'
2019-07-19 22:11:38 (1356): Status Report: CPU Time: '301.390625'
2019-07-19 23:58:14 (1356): Status Report: Elapsed Time: '24000.349737'
2019-07-19 23:58:14 (1356): Status Report: CPU Time: '415.859375'
2019-07-20 01:42:26 (1356): Status Report: Elapsed Time: '30000.511370'
2019-07-20 01:42:26 (1356): Status Report: CPU Time: '521.671875'
2019-07-20 03:28:43 (1356): Status Report: Elapsed Time: '36000.841295'
2019-07-20 03:28:43 (1356): Status Report: CPU Time: '636.000000'
2019-07-20 05:14:21 (1356): Status Report: Elapsed Time: '42001.020304'
2019-07-20 05:14:21 (1356): Status Report: CPU Time: '746.953125'
2019-07-20 06:59:44 (1356): Status Report: Elapsed Time: '48001.033857'
2019-07-20 06:59:44 (1356): Status Report: CPU Time: '866.421875'
2019-07-20 08:45:43 (1356): Status Report: Elapsed Time: '54001.272972'
2019-07-20 08:45:43 (1356): Status Report: CPU Time: '986.468750'
2019-07-20 10:30:31 (1356): Status Report: Elapsed Time: '60001.683926'
2019-07-20 10:30:31 (1356): Status Report: CPU Time: '1105.906250'
2019-07-20 12:15:31 (1356): Status Report: Elapsed Time: '66002.297750'
2019-07-20 12:15:31 (1356): Status Report: CPU Time: '1224.671875'
2019-07-20 14:00:45 (1356): Status Report: Elapsed Time: '72002.397228'
2019-07-20 14:00:45 (1356): Status Report: CPU Time: '1343.031250'
2019-07-20 15:44:16 (1356): Status Report: Elapsed Time: '78002.397228'
2019-07-20 15:44:16 (1356): Status Report: CPU Time: '1455.312500'
2019-07-20 17:26:00 (1356): Status Report: Elapsed Time: '84002.397228'
2019-07-20 17:26:00 (1356): Status Report: CPU Time: '1557.265625'
2019-07-20 19:08:23 (1356): Status Report: Elapsed Time: '90002.397228'
2019-07-20 19:08:23 (1356): Status Report: CPU Time: '1668.625000'

So, it must be a work done in the localhost. But RDP say only ===== Tasks ========
Have no idea which work is done.
Funny idea is a longrunner Sherpa ;-))
ID: 6464 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 8 Apr 15
Posts: 751
Credit: 11,608,083
RAC: 1,481
Message 6465 - Posted: 20 Jul 2019, 19:18:14 UTC

Good morning

My three 2-core tasks are still running 28% and close to 16 hours running.

(HAL 5000 is the one running these)

ID: 6465 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Ray Murray
Avatar

Send message
Joined: 13 Apr 15
Posts: 138
Credit: 2,945,852
RAC: 0
Message 6466 - Posted: 20 Jul 2019, 19:38:05 UTC

Maeax
I got a couple of
====Tasks====
lines when the guest started up but then the Tasks started and seem to be working fine since, returning and requesting new tasks. Perhaps your VM isn't able to call out? Perhaps, being Linux on Linux, it needs something different from Linux on Windows. Perhaps I've just been lucky that it appears to have worked for me first time.
I wouldn't expect any reply or update from Laurence until Monday. Maybe it would be better to release new apps early in the week rather than on a Friday.
ID: 6466 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Crystal Pellet
Volunteer tester

Send message
Joined: 13 Feb 15
Posts: 1180
Credit: 815,336
RAC: 238
Message 6467 - Posted: 20 Jul 2019, 19:38:23 UTC

Good evening Magic,

That task has created a Linux Scientific host for you, that should run the native task with an inside BOINC running,
but that host did not get any tasks:

https://lhcathomedev.cern.ch/lhcathome-dev/show_host_detail.php?hostid=3870
ID: 6467 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
Profile Magic Quantum Mechanic
Avatar

Send message
Joined: 8 Apr 15
Posts: 751
Credit: 11,608,083
RAC: 1,481
Message 6468 - Posted: 20 Jul 2019, 20:03:01 UTC - in response to Message 6467.  

Hi CP

Well we seem to get that here by running these tasks on a pc with a Windows OS
The IP address is 127.0.0.1

The 127.0.0.1 is the loopback Internet protocol (IP) address also referred to as the “localhost.”
The address is used to establish an IP connection to the same machine or computer being used by the end-user.

The tasks actually are running in the other pc HAL 5000
https://lhcathomedev.cern.ch/lhcathome-dev/results.php?hostid=3814
ID: 6468 · Rating: 0 · rate: Rate + / Rate - Report as offensive     Reply Quote
1 · 2 · 3 · Next

Message boards : General Discussion : The BOINC VM application


©2024 CERN