LHC@home

How can I avoid getting native tasks? I have SuSE Linux.

You can deselect "Run test applications" in your preferences to avoid getting native tasks.

Yes, but I would lose all sixtracktest. Atlas tasks and SixTrack are the only tasks not failing on my 3 computers, a Windows 10 PC with 22 GB RAM and two Linux boxen with SuSE Linux. I check all Atlas task results for the HITS file, Some tasks on the Windows PC are completed and validated but produce no HITS file On the Linux computers, since they have only 8 GB RAM I run single core tasks, .2 on the Windows 10 PC.
Tullio

i am getting more and more of these "195 (0x000000C3) EXIT_CHILD_FAILED" errors.

in the last 5 days there were 8 of these:
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158595002
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158560174
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158518636
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158509582
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158456958
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158447999
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158413919
https://lhcathome.cern.ch/lhcathome/result.php?resultid=158391109

what is wrong with the setup? most of the native tasks work flawlessly.

The startup script runs these commands to check CVMFS:

cvmfs_config probe
cvmfs_config stat atlas.cern.ch
cvmfs_config stat atlas-condb.cern.ch

The first command is not checked if it succeeds or fails. If the second or third command fails for any reason you get the "CVMFS not found" error. I don't know why exactly the connection to atlas.cern.ch would fail but the connection to atlas-condb.cern.ch succeeds.

There could be some clues in syslog. I get messages like these for all CVMFS repositories the app accesses:

Oct  3 17:08:23 mint cvmfs2: (atlas.cern.ch) failed to resolve IP addresses for ca-proxy.cern.ch (4 - unknown host name)
Oct  3 17:08:23 mint cvmfs2: (atlas.cern.ch) geographic order of servers retrieved from cernvmfs.gridpp.rl.ac.uk
Oct  3 17:08:32 mint cvmfs2: (atlas.cern.ch) CernVM-FS: linking /cvmfs/atlas.cern.ch to repository atlas.cern.ch

For completeness, the other checks the script runs are:

singularity --version
singularity exec -B /cvmfs /cvmfs/atlas.cern.ch/repo/images/singularity/x86_64-slc6.img hostname

These are run on non SL6 hosts. If they fail the error message is "Singularity is not installed" for the first one and "Singularity isnt working..." for the second.

thanks for the information and tips.
unfortunatly i havent seen any hints in the syslogs for why its failing sometimes. But i will look deeper into them once i have more time.

is anyone running successfully native tasks on ubuntu?
according to this post it should work:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4396&postid=32273

but this post tells a different story:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=4396&postid=32335

since these posts are rather old, and a new version has been release (2.52), does anyone have experience with running native tasks on ubuntu or other linux distros except then CentOS and SL?

is anyone running successfully native tasks on ubuntu?

They don't send any to my Ubuntu machine, so I can't test them. It appears that they simply are not ready for Ubuntu yet, but they have not made that clear.

Back when I was running the native app on Mint I had changed BOINC to report kernel version as 3.13.0-123-generic.not.really.sl6. I changed the kernel version back to what it really is two or three weeks ago and haven't seen any native tasks since then.

Back when I was running the native app on Mint I had changed BOINC to report kernel version as 3.13.0-123-generic.not.really.sl6. I changed the kernel version back to what it really is two or three weeks ago and haven't seen any native tasks since then.

If it worked OK on Mint, they should be able to get it to work on Ubuntu. Their silence on the subject is deafening. Have they given up on it?

Back when I was running the native app on Mint I had changed BOINC to report kernel version as 3.13.0-123-generic.not.really.sl6. I changed the kernel version back to what it really is two or three weeks ago and haven't seen any native tasks since then.

If it worked OK on Mint, they should be able to get it to work on Ubuntu. Their silence on the subject is deafening. Have they given up on it?

Definitely not! The silence is more due to the fact that it is working extremely well on the currently supported Linux flavours. One issue we still have is that suspending a WU doesn't seem to work, the WU keeps on running even if BOINC thinks it is suspended. On the server-like hosts with RHEL 6 or 7 we currently run on this isn't really a problem since the tasks run continuously in the background, but people with Ubuntu on their laptops for example probably need the suspend to work. It seems like there is a big appetite among Ubuntu users to try it so we'll put some effort into fixing the issue.

One issue we still have is that suspending a WU doesn't seem to work, the WU keeps on running even if BOINC thinks it is suspended. On the server-like hosts with RHEL 6 or 7 we currently run on this isn't really a problem since the tasks run continuously in the background, but people with Ubuntu on their laptops for example probably need the suspend to work. It seems like there is a big appetite among Ubuntu users to try it so we'll put some effort into fixing the issue.

OK, thanks for the update. I run my Ubuntu machines 24/7 and they are dedicated to BOINC work, so it should not be an issue for me. I would like the increased efficiency and whatever other benefits (smaller memory?) accrue when you get it working.

One of the other main benefits is that the CVMFS cache is shared among all the tasks, so you don't have to download as much data for each WU.

Out of interest I checked results from the last week and I was quite surprised that almost half come from native WU!

App version Total results
1.01 x86_64-apple-darwin [vbox64_mt_mcore_atlas] 1591
2.52 x86_64-pc-linux-gnu [native_mt] 54512
1.01 x86_64-pc-linux-gnu [vbox64_mt_mcore_atlas] 14063
1.01 windows_x86_64 [vbox64_mt_mcore_atlas] 56143

App version Total results
1.01 x86_64-apple-darwin [vbox64_mt_mcore_atlas] 1591
2.52 x86_64-pc-linux-gnu [native_mt] 54512
1.01 x86_64-pc-linux-gnu [vbox64_mt_mcore_atlas] 14063
1.01 windows_x86_64 [vbox64_mt_mcore_atlas] 56143

At the moment we have 120k used from 200k for the last tasks for Atlas!
It smelt down very quickly ;-)

Bonsoir !

I received my first ATLAS native app WUs yesterday but it failed with "CVMFS not found".
So, i installed CVMFS from CERN https://cernvm.cern.ch/portal/filesystem/downloads but it failed again.
Example : https://lhcathome.cern.ch/lhcathome/result.php?resultid=169250409

After that, i read this thread (should have start by that !) and saw there's an install script but it's for RH, i'm on Debian.
I modified it and now :

$ cvmfs_config probe
Probing /cvmfs/atlas.cern.ch... OK
Probing /cvmfs/atlas-condb.cern.ch... OK
Probing /cvmfs/grid.cern.ch... OK

Seems good !!! :)
I can share the script if you want/need it.

But now i've no more ATLAS native WU to test... Is there a way to test everything is good now before i receive a new one ? Is it possible to run the ATLAS binaries outsite BOINC client ?

Thanks !

I received my first ATLAS native app WUs yesterday but it failed with "CVMFS not found".

In addition to CVMFS, you need to install Singularity. This works for me on Ubuntu 16.04, with the latest version being Singularity 2.4.2:

First, check the latest version on GitHub: https://github.com/singularityware/singularity/releases
And substitute that for "$VERSION"
So $VERSION=2.4.2 (as of 6 December 2017)

wget https://github.com/singularityware/singularity/releases/download/$VERSION/singularity-$VERSION.tar.gz
=> wget https://github.com/singularityware/singularity/releases/download/2.4.2/singularity-2.4.2.tar.gz

tar xvf singularity-$VERSION.tar.gz
=> tar xvf singularity-2.4.2.tar.gz

cd singularity-$VERSION
=> cd singularity-2.4.2

./configure --prefix=/usr/local
make
sudo make install

To check the version installed: singularity --version
To check the usage: singularity --help

Then, I would reboot and try again.

Thanks Jim.
Yes, i installed Singularity too but from Debian repositories. The actual version is 2.3.2 on testing, is that version ok or do i need to backport from unstable (v2.4.1) ?

To test it, do i need to wait a new WU or is there a way to execute the ATLAS binaries as a "standalone" app ?

The actual version is 2.3.2 on testing, is that version ok or do i need to backport from unstable (v2.4.1) ?

To test it, do i need to wait a new WU or is there a way to execute the ATLAS binaries as a "standalone" app ?

I expect 2.3.2 is OK; there have been several upgrades recently that probably don't affect our use. But I think 2.4.2 is stable if you want to try it; I am having no problems with it. There should not be much difference between Debian and Ubuntu, though I am no expert on the subject. Probably the main issue is the Linux kernel; I am on 4.10.0.42.

I think if both CVMFS and Singularity check out OK on their tests, that should be enough. I don't know how to test it further on ATLAS except to download it.

Hello

New WU, new failure :/ https://lhcathome.cern.ch/lhcathome/result.php?resultid=169659259
But i made some progress ! CVMFS is ok now but Singularity terminated with an error :

Checking for CVMFS
CVMFS is installed
OS:cat: /etc/redhat-release: No such file or directory

This is not SLC6, need to run with Singularity....
Checking Singularity...
Singularity is installed
copy /var/lib/boinc-client/slots/3/shared/RTE.tar.gz
copy /var/lib/boinc-client/slots/3/shared/start_atlas.sh
copy /var/lib/boinc-client/slots/3/shared/input.tar.gz
copy /var/lib/boinc-client/slots/3/shared/ATLAS.root_0
export ATHENA_PROC_NUMBER=7;start atlas job with PandaID=3748950343
Testing the function of Singularity...

Singularity isnt working...

The script "ATLAS_run_atlas_2.52_x86_64-pc-linux-gnu" seems too failed with that command : "singularity exec -B /cvmfs /cvmfs/atlas.cern.ch/repo/images/singularity/x86_64-slc6.img hostname".
When i execute it on a terminal, i get that result :

WARNING: Not mounting requested bind point (already mounted in container): /cvmfs
Segmentation fault

Any idea ?

Auto-reply... Segfault could be solved by passing the variable "vsyscall=emulate" to the kernel. http://singularity.lbl.gov/faq#segfault-on-bootstrap-of-centos-image

Waiting for next WU to test...

I always wonder why everyone is having such a hard time setting up their systems for the native app. I used Boinc to get accustomed to Linux (Ubuntu in my case) and it worked like a charm (once I figured out why the scripts wouldn't run).

I used singularity-container from NeuroDebian repositories, if that helps.