ATLAS simulation V1.22 released


Advanced search

Message boards : News : ATLAS simulation V1.22 released

Author Message
Profile Wenjing Wu
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 23 Jun 14
Posts: 31
Credit: 2,849,678
RAC: 18
      
Message 290 - Posted: 16 Jul 2014, 3:48:40 UTC
Last modified: 16 Jul 2014, 3:49:52 UTC

V1.22 is released to change the virtualbox snapshot taking frequency from 1 hour to 15 minutes. If volunteers do not feel comfortable with this frequency, one can also change this value in their computing preferences via the website interface.

Matthias Lehmkuhl
Send message
Joined: 26 Jun 14
Posts: 12
Credit: 25,793
RAC: 2
      
Message 308 - Posted: 17 Jul 2014, 7:37:28 UTC

One question, which value must be changed for the virtualbox snapshot?

could just start one result, the download of the additional files is also in this version not possible behind a proxy
http://atlasathome.cern.ch/result.php?resultid=104373

Looks like the error
VT-x/AMD-V hardware acceleration is not available on your system. Your 64-bit guest will fail to detect a 64-bit CPU and will not be able to boot.
is also not solved.
On an other machine I saw that the VM was started, but then stopped. Will check it tonight CEST.
T4T (new -> vLHCathome) Theory Simulations v260.79 (vbox64) is running on that machine without problems, also do run RNA World results (vbox64)
I found, that this error happens on my machine, when the storage controller with the vdi image couldn't be added to the VM

This is the part from a successful configuration.
...
Disabling Drag and Drop Support for VM.
Adding storage controller to VM.
Adding virtual disk drive to VM. (vm_image.vdi)
Adding network bandwidth throttle group to VM. (Defaulting to 1024GB)
...
Starting VM.
...
the vdi-image could be added to the VM

And here the part from an unsuccessful configuration
...
Disabling Drag and Drop Support for VM.
Hardware acceleration failed with previous execution. Disabling VirtualBox hardware acceleration support.
ERROR: Invalid configuration. VM type requires acceleration but the current configuration cannot support it.
vboxwrapper (7.5.26094): starting
...
the vdi-image couldn't be added to the VM
If I manually add the storage controller and vdi image to the VM, the VM will start. Will also check this tonight.
____________
Matthias

Profile ritterm
Avatar
Send message
Joined: 22 Jun 14
Posts: 43
Credit: 286,157
RAC: 2
      
Message 309 - Posted: 17 Jul 2014, 13:47:43 UTC - in response to Message 290.
Last modified: 17 Jul 2014, 13:48:36 UTC

V1.22 is released to change the virtualbox snapshot taking frequency from 1 hour to 15 minutes. If volunteers do not feel comfortable with this frequency, one can also change this value in their computing preferences via the website interface.

Excuse my ignorance, but I'm not sure what the impact of this change is. What kinds of things do volunteers need to be worried about if snapshots are going to be more frequent?

Thanks,

MarkR
____________

dskagcommunity
Avatar
Send message
Joined: 11 Jul 14
Posts: 19
Credit: 44,513
RAC: 0
    
Message 310 - Posted: 17 Jul 2014, 15:13:55 UTC

I would say the HD/SSD Usage will grow exponential. The more units one machine run at the same time, the more snapshots must be saved at the same time and that could impact the Workunits/day Count too when it runs bad. Depending how much hd usage you already have and what HD/SSD Type. I think SSD user will not get that much impact because of its big IO and read/Write power (perhaps its getting faster "old"). But ATLAS already has the biggest IO Throughput from all projects i had running so far, so im "prepared" for that ^^

Matthias Lehmkuhl
Send message
Joined: 26 Jun 14
Posts: 12
Credit: 25,793
RAC: 2
      
Message 312 - Posted: 17 Jul 2014, 21:34:49 UTC

Could just check the storage controller settings for the downloaded results, they are missing again for all 4 downloaded results.

I only can set up one result with the virtual disk.
For the next storage controller I can't add the virtual disk -> reason is the same virtual disk UUID, this is not accepted by virtualbox system.
The virtual disk image is copied from the project dir to the slot dir, so the disk UUID is the same.

Please change the settings on server, so that only one result is send for a computer until this is reported.
This is done by vLHCathome and RNA World vbox too.
____________
Matthias

Profile Wenjing Wu
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 23 Jun 14
Posts: 31
Credit: 2,849,678
RAC: 18
      
Message 380 - Posted: 23 Jul 2014, 2:52:58 UTC - in response to Message 312.

Hi,
Thanks for the information!I have changed the total job for each host to be 1. It used to be 1 job for each processor.

Could just check the storage controller settings for the downloaded results, they are missing again for all 4 downloaded results.

I only can set up one result with the virtual disk.
For the next storage controller I can't add the virtual disk -> reason is the same virtual disk UUID, this is not accepted by virtualbox system.
The virtual disk image is copied from the project dir to the slot dir, so the disk UUID is the same.

Please change the settings on server, so that only one result is send for a computer until this is reported.
This is done by vLHCathome and RNA World vbox too.

Jacob Klein
Send message
Joined: 21 Jun 14
Posts: 48
Credit: 27,798
RAC: 0
    
Message 381 - Posted: 23 Jul 2014, 3:04:24 UTC

Fyi, I don't believe that RNA World limits to 1-per-host, but vLHC does. For ATLAS, I have 4 VMs running concurrently fine on one of my Windows 8.1 clients.

Could you please explain why you are changing ATLAS to 1-per-host? Also, is the long term plan to keep it 1-per-host, or will you eventually open things up again to allow many-per-host?

Regards,
Jacob

Tom*
Send message
Joined: 28 Jun 14
Posts: 118
Credit: 8,761,428
RAC: 215
      
Message 382 - Posted: 23 Jul 2014, 4:36:45 UTC

Goodbye sweet ATLAS

you know not what you have done.

markus tingsnaes
Send message
Joined: 4 Jul 14
Posts: 4
Credit: 77,834
RAC: 0
    
Message 383 - Posted: 23 Jul 2014, 4:46:27 UTC

WHY ??? is it a limit 1 wu per host now ??
I had 8 wu running at the same time on my computer since I started here and it works just fine.

Best regads

Profile Wenjing Wu
Project administrator
Project developer
Project tester
Project scientist
Avatar
Send message
Joined: 23 Jun 14
Posts: 31
Credit: 2,849,678
RAC: 18
      
Message 385 - Posted: 23 Jul 2014, 6:01:46 UTC - in response to Message 380.

After talking to the BOINC developer and following several follow up messages in this thread, it is confirmed that multiple virtual machines can be run on the same host without having the conflicting UUID problem. So I am changing the setup back to 1 job for each processor!
Sorry for the inconvenience!


Hi,
Thanks for the information!I have changed the total job for each host to be 1. It used to be 1 job for each processor.

Could just check the storage controller settings for the downloaded results, they are missing again for all 4 downloaded results.

I only can set up one result with the virtual disk.
For the next storage controller I can't add the virtual disk -> reason is the same virtual disk UUID, this is not accepted by virtualbox system.
The virtual disk image is copied from the project dir to the slot dir, so the disk UUID is the same.

Please change the settings on server, so that only one result is send for a computer until this is reported.
This is done by vLHCathome and RNA World vbox too.

____________

Jacob Klein
Send message
Joined: 21 Jun 14
Posts: 48
Credit: 27,798
RAC: 0
    
Message 397 - Posted: 23 Jul 2014, 10:43:58 UTC

Why have a limit at all? If I have a computer that only connects to the internet every 4 days, I expect to be supplied with 4 days of work for each processor.

That's how BOINC is supposed to work, anyway.

Please consider removing the restriction altogether.

Nick Name
Send message
Joined: 22 Jun 14
Posts: 13
Credit: 135,078
RAC: 0
    
Message 413 - Posted: 24 Jul 2014, 1:26:55 UTC

I use an app_config file to limit how many jobs run at the same time. Is there some reason that wouldn't work for those who can only run one at a time?

Profile Yeti
Avatar
Send message
Joined: 20 Jul 14
Posts: 699
Credit: 22,597,832
RAC: 211
      
Message 419 - Posted: 24 Jul 2014, 6:41:12 UTC

Could you post your App_config file here

Yeti

Crystal Pellet
Send message
Joined: 25 Jun 14
Posts: 39
Credit: 59,718
RAC: 0
    
Message 420 - Posted: 24 Jul 2014, 6:55:18 UTC - in response to Message 419.
Last modified: 24 Jul 2014, 6:56:00 UTC

Could you post your App_config file here

Yeti

I use this one to reduce the number of 'Running' ATLAS-tasks on my 8-threaded machine:

<app_config>
<app>
<name>ATLAS</name>
<max_concurrent>4</max_concurrent>
</app>
</app_config>

Nick Name
Send message
Joined: 22 Jun 14
Posts: 13
Credit: 135,078
RAC: 0
    
Message 435 - Posted: 24 Jul 2014, 18:02:31 UTC - in response to Message 420.

Could you post your App_config file here

Yeti

I use this one to reduce the number of 'Running' ATLAS-tasks on my 8-threaded machine:

<app_config>
<app>
<name>ATLAS</name>
<max_concurrent>4</max_concurrent>
</app>
</app_config>


I use the same, but set max_concurrent to 2.

Message boards : News : ATLAS simulation V1.22 released