LHC@home

I'm getting "No tasks are available for Theory Simulation" yet the server status says there are thousands.
It was working earlier today.

Seeing no problems ftm.

In another post you mentioned that you tried Theory native on Linux.
To get those you set your prefs to "Run native if available?".

This must be disabled if you connect Windows computers to the same venue.

In another post you mentioned that you tried Theory native on Linux.
To get those you set your prefs to "Run native if available?".

This must be disabled if you connect Windows computers to the same venue.

Ah! That explains it. I assumed since native isn't available for Windows, the tick would get ignored. Another bug.

Since I can't get them to work in Linux, I'll untick it for now. Any idea why they don't work in Linux? Do I need a certain library? It's a pretty bare bones Linux setup. Everything fails with a computation error in 2 seconds like something is missing. I assume native doesn't use VB, which is not installed.

Just read the logfiles and lots of forum posts explaining it.
What you do instead is to always complain about everything that you don't understand.


It's like a red button with "DON'T PRESS HERE!" written on it.
You press it and later you complain that "PRESS HERE" was written on it.

Do not comment this!
Instead, think about it!

I have searched, I can find no information on Linux and native tasks. I see no reason a native Linux task shouldn't just work, like any other non-VB project.

And you've yet to explain why the tasks can't keep a copy of commonly downloaded data sets. If Boinc can't do it, suggest it to them.

Unlike other users you always claim to be an expert for everything.
Now search again.

Hint:
local CVMFS client

I'm only an expert in 95% of the world.

I had never heard of CVFMS, so I didn't know to search for it. After finding info I'm none the wiser.

Hint: I'm a Windows user, anything in Linux is Greek to me and all I can do is enter commands as instructed. I could do DOS, but Linux is beyond my capabilities. I've used GUI only since Windows 95, and I'm surprised Linux hasn't moved on yet.

Then you focussed on the wrong 95 % and the remaining 5 % is what matters.

I've used GUI only since Windows 95, and I'm surprised Linux hasn't moved on yet.

Linux can be run via GUI.
I'm using it every day for many years.
It's obviously in the remaining 5 %.

Then you focussed on the wrong 95 % and the remaining 5 % is what matters.

I've used GUI only since Windows 95, and I'm surprised Linux hasn't moved on yet.

Linux can be run via GUI.
I'm using it every day for many years.
It's obviously in the remaining 5 %.

Nobody can know everything.

I am of course using a Linux GUI. XUbuntu 20. But the command line seems to be used for absolutely everything.

So how do I make it run native apps?

So how do I make it run native apps?

Check the native box in your https://lhcathome.cern.ch/lhcathome/prefs.php?subset=project

So how do I make it run native apps?

Check the native box in your https://lhcathome.cern.ch/lhcathome/prefs.php?subset=project

I did, but it downloaded 200 tasks and failed every one in 2 seconds. Over at CPDN, you need to get special libraries to run it (32 bit ones). I was thinking something similar here, Linux has everything missing and nothing auto installs like Windows.

I looked at all of your hosts and it looks like this is the error problem you have,
I must admit I never have that happen with Windows here or over at -dev

I see one with just about a dozen errors with this

Ncat: Connection to 137.138.55.253 failed: Connection timed out.
2023-08-08 07:36:04 (23724): Guest Log: Ncat: Trying next address...
2023-08-08 07:36:04 (23724): Guest Log: libnsock nsock_connect_tcp(): TCP connection requested to 2001:1458:d00:17::13:80 (IOD #1) EID 16
2023-08-08 07:36:04 (23724): Guest Log: libnsock nsock_trace_handler_callback(): Callback: CONNECT ERROR [Network is unreachable (101)] for EID 16 [2001:1458:d00:17::13:80]
2023-08-08 07:36:04 (23724): Guest Log: Ncat: Network is unreachable.

But the one with over 200 errors is that Linux Ubuntu and yes you tend to have to figure everything out and do file edits and the rest......if it was me I would run it on Windows 10 and give up on the Natives and all the rest BUT that is up to you of course.

<message>
process exited with code 195 (0xc3, -61)</message>
<stderr_txt>
04:04:48 (35142): wrapper (7.15.26016): starting
04:04:48 (35142): wrapper (7.15.26016): starting
04:04:48 (35142): wrapper: running ../../projects/lhcathome.cern.ch_lhcathome/cranky-0.0.32 ()
04:04:48 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Detected Theory App
04:04:48 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Checking CVMFS.
04:04:48 BST +01:00 2023-08-08: cranky-0.0.32: [ERROR] 'which' could not locate the command 'cvmfs_config'.

I have noticed many threads about this over the last 4 years and with Google even more

I looked at all of your hosts and it looks like this is the error problem you have,
I must admit I never have that happen with Windows here or over at -dev

I see one with just about a dozen errors with this

Ncat: Connection to 137.138.55.253 failed: Connection timed out.
2023-08-08 07:36:04 (23724): Guest Log: Ncat: Trying next address...
2023-08-08 07:36:04 (23724): Guest Log: libnsock nsock_connect_tcp(): TCP connection requested to 2001:1458:d00:17::13:80 (IOD #1) EID 16
2023-08-08 07:36:04 (23724): Guest Log: libnsock nsock_trace_handler_callback(): Callback: CONNECT ERROR [Network is unreachable (101)] for EID 16 [2001:1458:d00:17::13:80]
2023-08-08 07:36:04 (23724): Guest Log: Ncat: Network is unreachable.

That may be caused by one of two things:

1) The internet connection was overloaded at the time. I only have 32Mbit down/7Mbit up, and 10 computers. Running LHC is difficult, lots of file transfers. I get the feeling the programs can get impatient and give up instead of putting up with a slow connection.
2) I had the VPN switched on at the time. Might affect all the computers as they go through a Windows network bridge on the main computer which has the VPN.

But the one with over 200 errors is that Linux Ubuntu and yes you tend to have to figure everything out and do file edits and the rest......if it was me I would run it on Windows 10 and give up on the Natives and all the rest BUT that is up to you of course.

I wanted to try the natives and see if they were better than Oracle Virtualpox. Maybe they'd even have the sense to not download the same file multiple times on the same computer.

<message>
process exited with code 195 (0xc3, -61)</message>
<stderr_txt>
04:04:48 (35142): wrapper (7.15.26016): starting
04:04:48 (35142): wrapper (7.15.26016): starting
04:04:48 (35142): wrapper: running ../../projects/lhcathome.cern.ch_lhcathome/cranky-0.0.32 ()
04:04:48 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Detected Theory App
04:04:48 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Checking CVMFS.
04:04:48 BST +01:00 2023-08-08: cranky-0.0.32: [ERROR] 'which' could not locate the command 'cvmfs_config'.

I have noticed many threads about this over the last 4 years and with Google even more

Looks like I need this CVMFS thing. There must be a guide somewhere....

I got CVMFS installed, although the instructions only gave the ATLAS repositories (CVMFS_REPOSITORIES=atlas.cern.ch,atlas-condb.cern.ch,grid.cern.ch), I can't find the Theory ones.

I now get further on the task's attempt to start but it's still failing, can anyone interpret this?:

<core_client_version>7.18.1</core_client_version>
<![CDATA[
<message>
process exited with code 195 (0xc3, -61)</message>
<stderr_txt>
10:26:37 (129923): wrapper (7.15.26016): starting
10:26:37 (129923): wrapper (7.15.26016): starting
10:26:37 (129923): wrapper: running ../../projects/lhcathome.cern.ch_lhcathome/cranky-0.0.32 ()
10:26:37 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Detected Theory App
10:26:37 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Checking CVMFS.
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Checking runc.
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Creating the filesystem.
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Using /cvmfs/cernvm-prod.cern.ch/cvm3
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Updating config.json.
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Running Container 'runc'.
container_linux.go:336: starting container process caused "process_linux.go:293: applying cgroup configuration for process caused \"mountpoint for cgroup not found\""
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Container 'runc' finished with status code 1.
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Preparing output.
10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [ERROR] No output found.
10:26:38 (129923): cranky exited; CPU time 0.316500
10:26:38 (129923): app exit status: 0xce
10:26:38 (129923): called boinc_finish(195)

</stderr_txt>
]]>

As for CVMFS:

Apply the CVMFS settings mentioned here:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=6023&postid=48386
Then get the output from "cvmfs_config showconfig -s cvmfs-config" and "cvmfs_config showconfig -s atlas" and post them together with your:
/etc/cvmfs/default.local
/etc/cvmfs/domain.d/cern.ch.local
/etc/cvmfs/config.d/cvmfs-config.cern.ch.local



Mr P Hucker wrote:

Ncat: Connection to 137.138.55.253 failed: Connection timed out.
2023-08-08 07:36:04 (23724): Guest Log: Ncat: Trying next address...
2023-08-08 07:36:04 (23724): Guest Log: libnsock nsock_connect_tcp(): TCP connection requested to 2001:1458:d00:17::13:80 (IOD #1) EID 16
2023-08-08 07:36:04 (23724): Guest Log: libnsock nsock_trace_handler_callback(): Callback: CONNECT ERROR [Network is unreachable (101)] for EID 16 [2001:1458:d00:17::13:80]
2023-08-08 07:36:04 (23724): Guest Log: Ncat: Network is unreachable.

1) The internet connection was overloaded at the time. I only have 32Mbit down/7Mbit up, and 10 computers. Running LHC is difficult, lots of file transfers. I get the feeling the programs can get impatient and give up instead of putting up with a slow connection.
2) I had the VPN switched on at the time. Might affect all the computers as they go through a Windows network bridge on the main computer which has the VPN.

As for (1):
No.
Currently there are only Theory tasks available which hardly cause enough traffic to overload your internet connection.
Even if your connection is full with other traffic this would just cause a small delay.

As for (2):
Might be.

You should never run any ATLAS/Theory/CMS tasks via a VPN.
CERN implemented lots of measures to make the required data available as close as possible to your location (via Cloudflare's CDN).
A public VPN would route that traffic via the VPN server and a Cloudflare location close to the VPN server location.

In addition - depending on the VPN you are using - you may need to set up the correct port handling for non standard ports yourself.
Don't expect any help from the project in this area.

If you need a VPN for non BOINC/LHC traffic, use a separate computer on your side and do the VPN stuff only on that computer.

10:26:38 BST +01:00 2023-08-08: cranky-0.0.32: [INFO] Running Container 'runc'.
container_linux.go:336: starting container process caused "process_linux.go:293: applying cgroup configuration for process caused \"mountpoint for cgroup not found\""

As for this issue search the forum for cgroups v1/v2 and/or "systemd.unified_cgroup_hierarchy".


Finally:
Immediately and permanently stop your hidden or open innuendoes against any software, programmers or anybody else.
I will not repeat what you recently wrote as you know exactly what I mean.
Strictly respect the forum rules.
You insist in freedom of speech?
Well, then respect the limits!
Remember, you have been told that a couple of times over a couple of years.
And don't even think about any contradiction.

As for CVMFS:

Apply the CVMFS settings mentioned here:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=6023&postid=48386

I assume you mean:

In /etc/cvmfs/domain.d/cern.ch.local:
CVMFS_CONFIG_REPO_REQUIRED=yes
In /etc/cvmfs/config.d/cvmfs-config.cern.ch.local:
CVMFS_CONFIG_REPO_REQUIRED=no
Then run "sudo cvmfs_config reload".

I did that, except those files didn't exist, so I added the only line to them. Should the files have been there?

Then get the output from "cvmfs_config showconfig -s cvmfs-config"

I give up at this point. Getting Linux to map a network drive to Windows so I can get it copied across (it was a whole screenfull I'm not retyping). Impossible. Every single thing needs looking up, then a nest of several things under that, all requiring absurd commands to be typed in instead of just clicking stuff. It's high time everyone ditched the Mickey Mouse OS and went to Windows.

1) The internet connection was overloaded at the time. I only have 32Mbit down/7Mbit up, and 10 computers. Running LHC is difficult, lots of file transfers. I get the feeling the programs can get impatient and give up instead of putting up with a slow connection.
2) I had the VPN switched on at the time. Might affect all the computers as they go through a Windows network bridge on the main computer which has the VPN.

As for (1):
No.
Currently there are only Theory tasks available which hardly cause enough traffic to overload your internet connection.

CMS came up briefly and filled my buffer, CMS tasks were running.

Even if your connection is full with other traffic this would just cause a small delay.

You would think so, but I've seen tasks far more likely to give up if the network is busy.

As for (2):
Might be.

You should never run any ATLAS/Theory/CMS tasks via a VPN.
CERN implemented lots of measures to make the required data available as close as possible to your location (via Cloudflare's CDN).

A public VPN would route that traffic via the VPN server and a Cloudflare location close to the VPN server location.

In addition - depending on the VPN you are using - you may need to set up the correct port handling for non standard ports yourself.
Don't expect any help from the project in this area.

If you need a VPN for non BOINC/LHC traffic, use a separate computer on your side and do the VPN stuff only on that computer.

It's only occasionally, for accessing piratebay etc.

Finally:
Immediately and permanently stop your hidden or open innuendoes against any software, programmers or anybody else.
I will not repeat what you recently wrote as you know exactly what I mean.
Strictly respect the forum rules.
You insist in freedom of speech?
Well, then respect the limits!
Remember, you have been told that a couple of times over a couple of years.
And don't even think about any contradiction.

Look Stefanie, don't tell me what to do. I've not insulted anyone in here in the last few weeks, so I assume you're talking about the waste of bandwidth caused by not saving the huge data files locally so another task can use them. That cannot be labelled as anything other than gross stupidity. Why should CERN pay $$$ for bandwidth so I can download the same thing 100 times? And there are no limits to freedom of speech, or it wouldn't be called freedom.