Development for LHC@home

If you follow the link
https://lhcathomedev.cern.ch/lhcathome-dev/host_app_versions.php?hostid=37
on CP's computer detail page there are some strange entries.

You may compare the values "Max tasks per day = 502" for ATLAS Simulation 0.60 windows_x86_64 (vbox64_mt_mcore_atlas)
and "Max tasks per day = 9" for Theory Simulation 4.16 windows_x86_64 (vbox64_mt_mcore).

Are there other parameters that also have to be set?

If you follow the link
https://lhcathomedev.cern.ch/lhcathome-dev/host_app_versions.php?hostid=37
on CP's computer detail page there are some strange entries.

You may compare the values "Max tasks per day = 502" for ATLAS Simulation 0.60 windows_x86_64 (vbox64_mt_mcore_atlas)
and "Max tasks per day = 9" for Theory Simulation 4.16 windows_x86_64 (vbox64_mt_mcore).

Are there other parameters that also have to be set?

I reduced the max tasks per day for Theory to reduce the affect of blackhole hosts. Successful jobs +1 to this valued erros -1. I don't think this is and issue in this case.

Despite the limit ...

Do 21 Feb 2019 11:56:28 CET | lhcathome-dev | No tasks are available for Theory Simulation

Now I get "This computer has reached a limit on tasks in progress" with 'only' 7 tasks loaded.

I configured a host that simulates 2 CPUs.
It got 1 task and finished it.
Then it got 2 tasks (max) per RPC request until 5 tasks are in the buffer (2 running, 3 waiting)
Then "This computer has reached a limit on tasks in progress".

Buffer limits are 2d at the client and unlimited at the project page.

Now I get "This computer has reached a limit on tasks in progress" with 'only' 7 tasks loaded.

In the link from computezrmle I saw that max_jobs_per_day was less than n_jobs_today. I have updated max_jobs_per_day to be 50.

Now I get "This computer has reached a limit on tasks in progress" with 'only' 7 tasks loaded.

In the link from computezrmle I saw that max_jobs_per_day was less than n_jobs_today. I have updated max_jobs_per_day to be 50.

Like computezrmle I got the limit message from 7 tasks down to 4.
After having 'only' 4 in progress, I got a new task. So 5 tasks in progress seems now the limit.
The max tasks/day is increased by 1 with every valid task.

I configured a host that simulates 2 CPUs.
It got 1 task and finished it.
Then it got 2 tasks (max) per RPC request until 5 tasks are in the buffer (2 running, 3 waiting)
Then "This computer has reached a limit on tasks in progress".

Buffer limits are 2d at the client and unlimited at the project page.

I reconfigured the host to use 3 cores instead of 2.
The server doesn't send me more than 5 tasks in total.

Do 21 Feb 2019 14:35:24 CET | lhcathome-dev | This computer has reached a limit on tasks in progress

BTW:

Do 21 Feb 2019 14:35:24 CET | lhcathome-dev | No tasks are available for Theory Simulation

I reconfigured the host to use 3 cores instead of 2.
The server doesn't send me more than 5 tasks in total.

Do 21 Feb 2019 14:35:24 CET | lhcathome-dev | This computer has reached a limit on tasks in progress

BTW:

Do 21 Feb 2019 14:35:24 CET | lhcathome-dev | No tasks are available for Theory Simulation

You should get 6 now.

You should get 6 now.

Now I get up to 6 tasks on a 3 core host (2 tasks per RPC request).
3 running, 3 waiting.

I guess this is now how it is thought to work.

ScientificLinux 7-Vers.6 as VM in Virtualbox 5.2.26:

bash install_cvmfs_sin.sh Script from old Atlas-Server

default.local in /etc/cvmfs:
CVMFS_REPOSITORIES=cernvm-prod.cern.ch,grid.cern.ch,sft.cern.ch,alice.cern.ch
CVMFS_QUOTA_LIMIT=4096
CVMFS_CACHE_BASE=/scratch/cvmfs
sudo wget https://cern.ch/lfield/default.local -O /etc/cvmfs/default.local
CVMFS_HTTP_PROXY="DIRECT"

/etc/cvmfs/domain.d cern.local is a Copy from cern.ch.conf renamed to cern.local for openhtc.io:
CVMFS_SERVER_URL='http://s1cern-cvmfs.openhtc.io/cvmfs/@fqrn@;http://s1fnal-cvmfs.openhtc.io/cvmfs/@fqrn@;http://s1ral-cvmfs.openhtc.io/cvmfs/@fqrn@;http://s1bnl-cvmfs.openhtc.io/cvmfs/@fqrn@"

cvmfs_config reload

By default the maximum number of namespaces is set to 0. To fix this run:
echo 640 > /proc/sys/user/max_user_namespaces

Sorry for the short informations in this Checklist.

You should get 6 now.

Now I get up to 6 tasks on a 3 core host (2 tasks per RPC request).
3 running, 3 waiting.

I guess this is now how it is thought to work.

Maybe not yet?
I reduced the simulated cores from 3 to 2.
Now I still get 6 tasks until "This computer has reached a limit on tasks in progress".
When a task is reported the buffer fills up again to 6.

Seems that this setting overrules other parameters.

When a task is reported the buffer fills up again to 6.

My 8-thread buffer is always filled to 12 tasks since yesterday before have reached a limit on tasks in progress.

Sherpa 7:30 hours now. Is it possible to see RDP or other function in native App if Sherpa looping?
https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=2753887

Is there a limit for the task to break down?
Edit: Taken a deeper look in slot-Numbers of Boinc.
Have 7 tasks in parallel, but slot-Nr. are shown up to 21!
Maybe, they are not deleted after finishing?

Taken a deeper look in slot-Numbers of Boinc.
Have 7 tasks in parallel, but slot-Nr. are shown up to 21!
Maybe, they are not deleted after finishing?

My client (version 7.8.4) doesn't show this issue.
Your's is version 7.5.1.

Maybe a BOINC client issue?

Have Boinc 7.5.1 for Atlas-native also, without it.
Edit: Now 8:30 hours for this Sherpa!
In Atlas it was in the past also, I remember.
After Slot-Nr. 100 no more tasks possible.
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4396&postid=34374#34374

Sherpa 7:30 hours now. Is it possible to see RDP or other function in native App if Sherpa looping?
https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=2753887

You can monitor the process in a terminal with the command:

tail -F /var/lib/boinc-client/slots/0/cernvm/shared/runRivet.log

You probably have to adjust the command-line for the right slot-number.

Crystal thank you,
you have someone with a longrunner without success of Sherpa found ;-)
Display update finished (0 histograms, 0 events).
Updating display...
Display update finished (0 histograms, 0 events).
Channel_Elements::CheckMasses(): Strong deviation in masses
s2,p2: 1.24327e+07;(4503.96,1105.49,481.365,-2529.64) -> 1.24327e+07 : 0.0093001, rel = 2.06487e-06.
Updating display...
Display update finished (0 histograms, 0 events).
Updating display...
Display update finished (0 histograms, 0 events).
Updating display...

Please no restart... max. is three for this task:
https://lhcathomedev.cern.ch/lhcathome-dev/workunit.php?wuid=1877528

There's something wrong regarding the runtime calculation.


https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=2754021
Reported runtime 9 min 2 s?

14:41:15 (55252): wrapper (7.7.26015): starting
.
.
.
14:46:29 (55252): called boinc_finish(0)

https://lhcathomedev.cern.ch/lhcathome-dev/result.php?resultid=2753971
Reported runtime 1 h 2 min 1 s?

12:38:34 (6107): wrapper (7.7.26015): starting
.
.
.
13:26:59 (6107): called boinc_finish(0)

There's something wrong regarding the runtime calculation.

Even when running single-core tasks, the cpu-time is ever (mostly) higher than the elapsed time.
When you have idle cores the application is stealing from the free core(s).

See my Linux tasks and my Windows Vbox tasks.