LHC@home

Mirror for CentOS8 is out of Service since 22/01/31.
CVMFS is not supported for CentOS9 (only CentOS7 and CentOS8).

A second Squid is now running - CentOS9-VM (RedHat).
First Squid in CentOS8-VM (RedHat).

Edit: Atlas stopped upload, when Squid was changed during running.
After deselect of Squid, upload finished correct.

Mirror for CentOS8 is out of Service since 22/01/31.

Isn't that part of the change from CentOS to CentOS Stream? This has caused some "discussion" in the community, including proposals for LHC to move away from CentOS to some other distro. I don't know if there's any clear route on this yet.

CVMFS is not supported for CentOS9 (only CentOS7 and CentOS8).

May well be low-priority due to above.

---

Have two VM's with update from CentOS8 to CentOs8 Stream during the last months from the Distro.
CentOS is the public RHEL(RedHat-Enterprise Linux). RedHat is bought from IBM.

This Computer running CMS
(Download/Upload) ok with CentOS9(Stream)-VM RedHat squid:
https://lhcathome.cern.ch/lhcathome/results.php?hostid=10593998

This Computer running Atlas (Download is ok / Upload stuck):
https://lhcathome.cern.ch/lhcathome/results.php?hostid=10631979

This are the Boinc-Messages for upload:
21.03.2022 14:15:02 | LHC@home | Temporarily failed upload of 82fMDmnV7p0n9Rq4apoT9bVoABFKDmABFKDmKVOXDmABFKDm7i646m_0_r1956858398_ATLAS_hits: transient HTTP error
21.03.2022 14:15:02 | LHC@home | Backing off 00:02:22 on upload of 82fMDmnV7p0n9Rq4apoT9bVoABFKDmABFKDmKVOXDmABFKDm7i646m_0_r1956858398_ATLAS_hits
21.03.2022 14:15:03 | | Internet access OK - project servers may be temporarily down.
21.03.2022 14:15:03 | LHC@home | Temporarily failed upload of 82fMDmnV7p0n9Rq4apoT9bVoABFKDmABFKDmKVOXDmABFKDm7i646m_0_r1956858398_ATLAS_result: transient HTTP error

Have running one CMS on this PC, to see what happens with upload.
The other PC with only CMS, no problems.

Have running one CMS on this PC, to see what happens with upload.
The other PC with only CMS, no problems.

CMS-Task finished successful.
Explanation in Wikipedia for transient HTTP-Error is from a IBM360 Mainframe.
My first knowing is 370/153 to work on.

For Windows one squid CentOs9-RedHat for CMS and one squid CentOS8-RedHat for Atlas now installed and running well.

Have squid-RedHat complements:
# This clause only supports fast acl types. port 4827
#htcp_access deny all
htcp_access allow localnet
htcp_port 4827

Normal clients don't use HTCP.
It is used to exchange data between different squids.

Missing:
The configuration line(s) telling squid to query a sibling.
Without that HTCP will never be used.

Seeing in access.log lines from VM for this using.
cache.log show active HTCP in CentOS9 squid.
Have it also activated in CentOS8 squid!

cache.log show active HTCP in CentOS9 squid.

Of course it does. It tells you this squid listens to HTCP.

Seeing in access.log lines from VM for this using.

Post examples of that logfile lines.


And again:
If you configure squid1 to listen to HTCP you must tell squid2 to query squid1 via HTCP.
The options you posted so far do only configure the listener part of the connection.
Where is your sender part?

Post examples of that logfile lines.

Built it yourself, when you don't trust me!!!

I guess you can't because you don't know what you are doing.

I guess you can't because you don't know what you are doing.

2022/03/22 06:28:22 kid1| Accepting HTTP Socket connections at conn3 local=10.xxx.yyy.zz:3128 remote=[::] FD 11 flags=9
2022/03/22 06:28:22 kid1| Accepting HTCP messages on [::]:4827
2022/03/22 06:28:22 kid1| Sending HTCP messages from [::]:4827

2022/03/22 06:28:22 kid1| Accepting HTCP messages on [::]:4827
2022/03/22 06:28:22 kid1| Sending HTCP messages from [::]:4827

Those lines are from cache.log.
I already confirmed that they will appear as they tell you this squid (squid1) is able to communicate via HTCP (in both directions!).


Still missing are the lines from access.log you mentioned.
Also missing: the configuration lines telling squid2 to query a sibling (squid1 in this case).
The latter are a must as without them squid2 will never send a query to squid1.

Hint:
http://www.squid-cache.org/Doc/config/cache_peer/

no comment anymore for your answers in this thread!
http://www.squid-cache.org/Doc/config/htcp_access/

"Die beste Idee nützt niemandem,
wenn sie nicht verwertet wird." (Walter A. Heiby)

Please see the following page for official CERN information on CentOS versions: https://linux.web.cern.ch/centos/

Many software packages including CVMFS are not yet ready for CentOS9 but support for CVMFS is planned soon.

Thank you David.
This are good News to look in the future.
We can make a cut for Scientific Linux.
CentOS9 (Stream) works for me atm only for using Squid(RedHat) and
waiting for using CVMFS for Atlas-native and Theory-native with Boinc.

Please see the following page for official CERN information on CentOS versions: https://linux.web.cern.ch/centos/

Many software packages including CVMFS are not yet ready for CentOS9 but support for CVMFS is planned soon.

Is there a timewindow for using cvmfs and Boinc in CentOS9?

Please see the following page for official CERN information on CentOS versions: https://linux.web.cern.ch/centos/

Many software packages including CVMFS are not yet ready for CentOS9 but support for CVMFS is planned soon.

Is there a timewindow for using cvmfs and Boinc in CentOS9?

I tried this today and ATLAS tasks worked fine out of the box. CVMFS was installed from the usual CVMFS repo and boinc-client from EPEL next.