LHC@home

This is a discussion thread to post comments and questions regarding the CVMFS Configuration used by LHC@home:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=5594

A previous version of the HowTo and older comments can be found here:
https://lhcathome.cern.ch/lhcathome/forum_thread.php?id=5342

Hello,

My systems are all configured to use at least 4GB of local CVMFS cache and according to stats, they seem to reach pretty high hit ratios.
Here is just a look at a couple of them:


Would a proxy still help here ?

Also are there specific Squid options one should set in regards to caching CVMFS data (ie. max object size, retention or refresh policy) ?

My systems ... seem to reach pretty high hit ratios.
Would a proxy still help here ?

Yes.
Here are 2 major reasons.

Each CVMFS client does only serve tasks that are running on the same box (or inside the same VM).
A single Squid serves all boxes and all VMs running at your site.


Tasks like ATLAS or CMS heavily use CERN's Frontier system.
Frontier requests data via HTTP but unlike CVMFS it has no own local cache.
A local Squid closes this gap and serves most Frontier requests from it's cache.

Also are there specific Squid options one should set in regards to caching CVMFS data (ie. max object size, retention or refresh policy) ?

It's all covered by the squid.conf in the Squid HowTo.

Some of Squid's original settings have been made decades ago and focus on surfing the web via slow connections.
The suggestions in this forum extend the original settings and are based on experience and analysing the data flow created by LHC@home.
Nonetheless, surfing arbitrary internet pages with this settings is still possible but since most of them use HTTPS the hitrates for them would drop to 0 %.
Questions regarding specific Squid options should be asked in the Squid thread.

https://cvmfs.readthedocs.io/en/stable/cpt-squid.html
Container and CVMFS:
https://cvmfs.readthedocs.io/en/stable/cpt-containers.html

Very nice; thanks.
But I think it should be pointed out that the automatic configuration download no longer applies, insofar as I can see.
(sudo wget https://lhcathome.cern.ch/lhcathome/download/default.local -O /etc/cvmfs/default.local)

Maybe it could be updated?

Very nice; thanks.
But I think it should be pointed out that the automatic configuration download no longer applies, insofar as I can see.
(sudo wget https://lhcathome.cern.ch/lhcathome/download/default.local -O /etc/cvmfs/default.local)

Maybe it could be updated?

I had this problem, as well. Probing immediately failed.
Perhaps that file could be updated with the minimum needed configuration, although I'm still unclear how one can actually optimize their configuration if it is just 1 or 2 machines on the same connection.

Perhaps that file could be updated with the minimum needed configuration ...

The file on the server is already up to date.
Be aware that it includes 2 optional settings (with proxy/without proxy) and one of them has to be activated by the user.

In general:
Native apps require more settings to be done by the user.
This is easier, faster and more reliable than to guess certain values.
In addition some steps require to be done by root.

although I'm still unclear how one can actually optimize their configuration

The simple answer

Cache as much as possible as close as possible to the point were it is used.
To avoid less efficient effort focus on the major bottlenecks first.


More LHC@home specific

CVMFS is heavily used but has it's own cache - one cache instance per machine.
A machine can't share it's CVMFS cache with other machines.
Each VM counts as individual machine.
Outdated or missing data is requested from the project servers.

Frontier is heavily used by ATLAS and CMS. It has no own local cache.
Each app sends all Frontier requests to the project servers.


Cloudflare's openhtc.io infrastructure helps to distribute CVMFS and Frontier data.
They run a very fast worldwide network and one of their proxy caches will most likely be located much closer to your clients than any project server.

VBox apps use openhtc.io by default but users running native apps have to set "CVMFS_USE_CDN=yes" in their CVMFS configuration.
This is disabled in the default configuration because lots of computers in various datacenters use special connections and require this to be set "OFF".


A local HTTP proxy closes the gap between openhtc.io and the local clients.
It can cache data for all local CVMFS and Frontier clients as well as offload openhtc.io and the project servers.

Perhaps that file could be updated with the minimum needed configuration ...

The file on the server is already up to date.
Be aware that it includes 2 optional settings (with proxy/without proxy) and one of them has to be activated by the user.

In general:
Native apps require more settings to be done by the user.
This is easier, faster and more reliable than to guess certain values.
In addition some steps require to be done by root.

although I'm still unclear how one can actually optimize their configuration

The simple answer

Cache as much as possible as close as possible to the point were it is used.
To avoid less efficient effort focus on the major bottlenecks first.


More LHC@home specific

CVMFS is heavily used but has it's own cache - one cache instance per machine.
A machine can't share it's CVMFS cache with other machines.
Each VM counts as individual machine.
Outdated or missing data is requested from the project servers.

Frontier is heavily used by ATLAS and CMS. It has no own local cache.
Each app sends all Frontier requests to the project servers.


Cloudflare's openhtc.io infrastructure helps to distribute CVMFS and Frontier data.
They run a very fast worldwide network and one of their proxy caches will most likely be located much closer to your clients than any project server.

VBox apps use openhtc.io by default but users running native apps have to set "CVMFS_USE_CDN=yes" in their CVMFS configuration.
This is disabled in the default configuration because lots of computers in various datacenters use special connections and require this to be set "OFF".


A local HTTP proxy closes the gap between openhtc.io and the local clients.
It can cache data for all local CVMFS and Frontier clients as well as offload openhtc.io and the project servers.

How does one go about cacheing as much as possible?
Not sure what happened in my case, then, since as soon as I downloaded
https://lhcathome.cern.ch/lhcathome/download/default.local -O /etc/cvmfs/default.local
I got immediate failures after probing.
Running the listed items in the how to fixed my issues, and I believe I also added the line containing openhtc.io.

Thank you for the help and excellent clarification.

VBox apps use openhtc.io by default but users running native apps have to set "CVMFS_USE_CDN=yes" in their CVMFS configuration.
This is disabled in the default configuration because lots of computers in various datacenters use special connections and require this to be set "OFF".

Release Notes from CVMFS-Documentation 2.7.5
2.1 Release Notes for CernVM-FS 2.7.5CernVM-FS 2.7.5 is a patch release.
It contains several bugfixes for the client.As with previous releases, upgrading clients should be seamless just by installing the new package from therepository. As usual, we recommend to update only a few worker nodes first and gradually ramp up once the newversion proves to work correctly.
Please take special care when upgrading a cvmfs client in NFS mode.Stratum 0 and stratum 1 servers do not necessarily need to update from version 2.7.4.2.1.1
Bug Fixes and Improvements•
[client] fix rare crash when kernel meta-data caches operate close to 4GB (CVM-1918)•
[client] let mount helper detect when CVMFS_HTTP_PROXY is defined but empty•
[client] add CVMFS_CLIENT_PROFILE and CVMFS_USE_CDN to the list of known parameters in cvmfs_config

Atlas-Applet in Windows is using CVMFS 2.6.3.

Atlas-Applet in Windows is using CVMFS 2.6.3.

CVMFS_USE_CDN makes it easier to switch between the traditional CVMFS server list and the Cloudflare server list.
Older setups had to configure this manually which is still possible.
It's all fine as long as an application from this project uses Cloudflare servers.

Even CMS VMs that use v2.4.4.0 work fine.
Related to CVMFS_USE_CDN it's more important to use a recent cvmfs-config-default package than to upgrade the CVMFS client:
http://ecsft.cern.ch/dist/cvmfs/cvmfs-config/

What does this mean???

cvmfs_config stat
/usr/bin/cvmfs_config: line 907: cd: /cvmfs/atlas.cern.ch: Transport endpoint is not connected

I have an old native app setup. I did the steps in the Howto V2 today. It didn't set up the CDN until I also installed the latest CVMFS config.

Also, will any caching proxy do? I set up an old copy of Polipo.

---

Found an ATLAS native log of a task that succeeded:

[2022-06-09 19:17:48] Checking for CVMFS
[2022-06-09 19:17:48] Probing /cvmfs/atlas.cern.ch... OK
[2022-06-09 19:17:48] Probing /cvmfs/atlas-condb.cern.ch... OK
[2022-06-09 19:17:48] Running cvmfs_config stat atlas.cern.ch
[2022-06-09 19:17:48] VERSION PID UPTIME(M) MEM(K) REVISION EXPIRES(M) NOCATALOGS CACHEUSE(K) CACHEMAX(K) NOFDUSE NOFDMAX NOIOERR NOOPEN HITRATE(%) RX(K) SPEED(K/S) HOST PROXY ONLINE
[2022-06-09 19:17:48] 2.7.1.0 8420 286 49476 105303 0 78 2841033 4194305 786 65024 0 238387 99.9362 703 40 http://cvmfs-s1fnal.opensciencegrid.org/cvmfs/atlas.cern.ch http://127.0.0.1:8123 1
[2022-06-09 19:17:48] CVMFS is ok
[2022-06-09 19:17:48] Efficiency of ATLAS tasks can be improved by the following measure(s):
[2022-06-09 19:17:48] The CVMFS client on this computer should be configured to use Cloudflare's openhtc.io.
[2022-06-09 19:17:48] Further information can be found at the LHC@home message board.

Nonetheless, there are some points that should be changed:

1.
"CVMFS_USE_CDN=yes" should be set in /etc/cvmfs/default.local.

Prior to "cvmfs_config reload" you can show the new configuration if you run:

cvmfs_config showconfig -s atlas.cern.ch |grep CVMFS_SERVER_URL

It should return a list of "*.openhtc.io" servers instead of the stratum-ones.



2.
127.0.0.1 should not be used as proxy IP as it allows only processes from the same box to access that network IP.
That's why CVMFS can contact a proxy on the same box but another box in your network can't, not even a VM on the proxy box.

Instead, configure your proxy to listen to the LAN IP of the box (e.g. 192.168.x.y) and configure your clients to connect to the proxy via that IP.

Also, will any caching proxy do? I set up an old copy of Polipo.

In principle, yes.
CVMFS talks HTTP, hence each HTTP proxy should be able to handle the requests.
Your log mentioned above shows that a proxy is used for CVMFS.

But:
The Squid configuration given in this forum contains some efficiency settings, e.g. cache large files except ATLAS EVNT files, and I don't know if they can be used for other proxies.

I just installed cvmfs on my Ubuntu machines but I am still getting an error.

Before I installed cvmfs I got this error:

<core_client_version>7.16.6</core_client_version>
<![CDATA[
<message>
process exited with code 195 (0xc3, -61)</message>
<stderr_txt>
22:31:22 (1258264): wrapper (7.7.26015): starting
22:31:22 (1258264): wrapper: running run_atlas (--nthreads 12)
[2022-10-24 22:31:22] Arguments: --nthreads 12
[2022-10-24 22:31:22] Threads: 12
[2022-10-24 22:31:22] Checking for CVMFS
[2022-10-24 22:31:22] No cvmfs_config command found, will try listing directly
[2022-10-24 22:31:22] ls: cannot access '/cvmfs/atlas.cern.ch/repo/sw': No such file or directory
[2022-10-24 22:31:22] Failed to list /cvmfs/atlas.cern.ch/repo/sw
[2022-10-24 22:31:22] ** It looks like CVMFS is not installed on this host.
[2022-10-24 22:31:22] ** CVMFS is required to run ATLAS native tasks and can be installed following https://cvmfs.readthedocs.io/en/stable/cpt-quickstart.html
[2022-10-24 22:31:22] ** and setting 'CVMFS_REPOSITORIES=atlas.cern.ch,atlas-condb.cern.ch' in /etc/cvmfs/default.local
22:41:23 (1258264): run_atlas exited; CPU time 0.010115
22:41:23 (1258264): app exit status: 0x1
22:41:23 (1258264): called boinc_finish(195)

So I installed cvmfs and checked the install and did not get any errors when I used

 cvmfs_config probe 

I also configured cvmfs as per the web page with

sed -i 's%#+dir:/etc/auto.master.d%+dir:/etc/auto.master.d%' /etc/auto.master
systemctl restart autofs

I downloaded a new task https://lhcathome.cern.ch/lhcathome/result.php?resultid=367523657

After it ran I am still getting the error:

<core_client_version>7.16.6</core_client_version>
<![CDATA[
<message>
process exited with code 195 (0xc3, -61)</message>
<stderr_txt>
03:03:32 (6989): wrapper (7.7.26015): starting
03:03:32 (6989): wrapper: running run_atlas (--nthreads 12)
[2022-10-25 03:03:32] Arguments: --nthreads 12
[2022-10-25 03:03:32] Threads: 12
2022-10-25 03:03:32] Checking for CVMFS
[2022-10-25 03:03:32] Probing /cvmfs/atlas.cern.ch... Failed!
[2022-10-25 03:03:32] Probing /cvmfs/atlas-condb.cern.ch... Failed!
[2022-10-25 03:03:32] cvmfs_config probe atlas.cern.ch atlas-condb.cern.ch failed, aborting the job
03:13:33 (6989): run_atlas exited; CPU time 0.163581
03:13:33 (6989): app exit status: 0x1
03:13:33 (6989): called boinc_finish(195)

---

Proud Founder and member of BOINC@AUSTRALIA



Have a look at my WebCam

Did you also install the latest config package?
It can be found here:
http://ecsft.cern.ch/dist/cvmfs/cvmfs-config/
Download the one matching your OS and starting with "cvmfs-config-default-latest...".

Then modify/create /etc/cvmfs/default.local according to the advice here and in this forum.
If unsure, post that file.


Run "cvmfs_config probe" first from a console.
As long as this fails it makes no sense to download any ATLAS/Theory native task.
This just results in lots of downloads and failed tasks.

I don't know which file to download from http://ecsft.cern.ch/dist/cvmfs/cvmfs-config/ I have Linux Ubuntu 20.04 5.4.0-131-generic #147-Ubuntu SMP Fri Oct 14 17:07:22 UTC 2022 x86_64 x86_64 x86_64 GNU/Linux



OK I started from scratch and did the following:

cd ~/
wget https://ecsft.cern.ch/dist/cvmfs/cvmfs-release/cvmfs-release-latest_all.deb
sudo dpkg -i cvmfs-release-latest_all.deb
rm -f cvmfs-release-latest_all.deb
sudo apt-get update
sudo apt-get install cvmfs cvmfs-config-default
sudo cvmfs_config setup
sudo wget https://lhcathome.cern.ch/lhcathome/download/default.local -O /etc/cvmfs/default.local
sudo cvmfs_config reload
sudo sed -i '$ a\kernel.unprivileged_userns_clone = 1' /etc/sysctl.conf
sudo sysctl -p
sudo wget http://lhcathome.cern.ch/lhcathome/download/create-boinc-cgroup -O /sbin/create-boinc-cgroup
sudo wget http://lhcathome.cern.ch/lhcathome/download/boinc-client.service -O /etc/systemd/system/boinc-client.service
sudo systemctl daemon-reload
sudo systemctl restart boinc-client

I downloaded a new task https://lhcathome.cern.ch/lhcathome/result.php?resultid=367712964 and it is now I still get the error

[2022-10-25 05:44:27] Threads: 12
[2022-10-25 05:44:27] Checking for CVMFS
[2022-10-25 05:44:28] Probing /cvmfs/atlas.cern.ch... Failed!
[2022-10-25 05:44:28] Probing /cvmfs/atlas-condb.cern.ch... Failed!
[2022-10-25 05:44:28] cvmfs_config probe atlas.cern.ch atlas-condb.cern.ch failed, aborting the job

What command do I need to run ??

---

Proud Founder and member of BOINC@AUSTRALIA



Have a look at my WebCam

I downloaded a new task ...

As already mentioned
This makes no sense as long as you run "cvmfs_config probe" in a console window and get this output:

Probing /cvmfs/atlas.cern.ch... Failed!

You just burn all native tasks as long as your CVMFS issue is not fixed!

I don't know which file to download ...

Since you mentioned "Ubuntu" this one should be fine:
http://ecsft.cern.ch/dist/cvmfs/cvmfs-config/cvmfs-config-default_latest_all.deb
Install it after or together with the main CVMFS client package, then run "cvmfs_config setup".

In your /etc/cvmfs/default.local add/uncomment this line:
CVMFS_HTTP_PROXY="auto;DIRECT"
Be aware of the quoting. It must be double quotes here because of the ";" within the value.

CVMFS requires user namespaces enabled.
How to do this depends on the linux kernel you are running.

This is a quick check:
sudo sysctl user.max_user_namespaces
Must not be "0"

Other hints can be found in the internet or here:
https://lhcathomedev.cern.ch/lhcathome-dev/forum_thread.php?id=486&postid=6520


Reboot after you made the changes.




create-boinc-cgroup and boinc-client.service don't play a role for now.
First focus on the CVMFS configuration.

Just too much trouble for an old fart like me. I will just run them in VirtualBox.

Thanks for trying to help me though.

Has anyone seen a scenario where running cvmfs_config probe returns success but native ATLAS fails due to such a probe failure when ran by the wrapper at task startup? What could be some possible reasons?
For context, WSL2 now supports systemd. I enabled it to see how things go with the hope of being able to run ATLAS multi-core on WSL2. It didn't go well for ATLAS as the probe failed according to the error log, however, it succeeds when I run it in isolation from the command prompt.