Hi Alastair,
we mainly use the cvmfs default configuration, but we run a lot more jobs that use opensciencegrid cvmfs repositories. For example darkside.opensciencegrid.org:
# cvmfs_config showconfig -s darkside.opensciencegrid.org | grep SERVER_URL
CVMFS_SERVER_URL='http://cvmfs-egi.gridpp.rl.ac.uk:8000/cvmfs/darkside.opensciencegrid.org;http://klei.nikhef.nl:8000/cvmfs/darkside.opensciencegrid.org;http://cvmfs-s1bnl.opensciencegrid.org:8000/cvmfs/darkside.opensciencegrid.org;http://cvmfs-s1fnal.opensciencegrid.org:8000/cvmfs/darkside.opensciencegrid.org;http://cvmfsrep.grid.sinica.edu.tw:8000/cvmfs/darkside.opensciencegrid.org' # from /cvmfs/config-egi.egi.eu/etc/cvmfs/domain.d/opensciencegrid.org.conf
It should use RAL as primary server, but if I do
curl "http://cvmfs-egi.gridpp.rl.ac.uk:8000/cvmfs/darkside.opensciencegrid.org/.cvmfspublished"
I get a 404, so RAL is probably not mirroring those repositories. The same is true for nikhef, the first available server in the list is the BNL server.
Cheers,
Robert
On 14/08/18 22:22, Alastair Dewhurst wrote:
> Hi all
>
> I received the message below earlier today.
>
> Manchester and Oxford appear to be using the BNL CVMFS Stratum 1. Its not breaking anything their end but he wondered if this was the intended configuration. Interestingly you have an awful lot of hits but not much bandwidth which may or may not be important. From the RAL Tier-1 point of view, we would certainly like to know if the CVMFS Stratum 0/1 services we offer are not providing a reliable service or missing features.
>
> Thanks
>
> Alastair
>
>
> Begin forwarded message:
>
> From: "John S. De Stefano Jr." <[log in to unmask]<mailto:[log in to unmask]>>
> Subject: CVMFS Squids in Europe
> Date: 14 August 2018 at 20:28:06 BST
> To: Saul Youssef <[log in to unmask]<mailto:[log in to unmask]>>
> Resent-From: "John S. De Stefano Jr." <[log in to unmask]<mailto:[log in to unmask]>>
> Resent-To: Alastair Dewhurst <[log in to unmask]<mailto:[log in to unmask]>>
>
> Hi Alastair,
>
> It happens that three of the 10 heaviest consumers of the BNL CVMFS Stratum One for this calendar year thus far have come from the UK **, with another coming from INFN:
>
> Hosts : Pages Hits Bandwidth Last visit
>
> squid2.cmsaf.mit.edu<http://squid2.cmsaf.mit.edu> 42427252 42427254 535.28 GB 14 Aug 2018 - 14:53
> squid.cmsaf.mit.edu 41566822 41566823 546.58 GB 14 Aug 2018 - 14:53
> its-condor-squid1.syr.edu 25237476 25237476 384.52 GB 14 Aug 2018 - 14:53
> squid1.cmsaf.mit.edu 17464328 17464328 459.09 GB 14 Aug 2018 - 14:53
> gryphn.phys.uconn.edu 13596860 13596860 2526.00 GB 14 Aug 2018 - 14:53
> t2software03.physics.ox.ac.uk ** 8945135 8945135 43.43 GB 14 Aug 2018 - 14:53
> vm55.in.tier2.hep.manchester.ac.uk ** 8917684 8917684 94.36 GB 14 Aug 2018 - 14:53
> cvmfs-s1a.sdcc.bnl.gov 8907863 8907863 1054.63 GB 14 Aug 2018 - 14:50
> vm53.in.tier2.hep.manchester.ac.uk ** 8451309 8451309 80.69 GB 14 Aug 2018 - 14:20
> squid-lhc-02.cr.cnaf.infn.it 7803641 7803641 12.74 GB 14 Aug 2018 - 14:53
>
> There is nothing inherently nor necessarily wrong with this, and it’s not causing any trouble. But with data and bandwidth rates this high, I wanted to check and ensure the Squids and local Strata One are healthy and not running out of resources (disk, memory, inodes). This could also be due to a change in access pattern, or just improved hardware and bandwidth on our side (or at both ends).
>
> Thanks for looking.
>
> ~John
>
>
> ########################################################################
>
> To unsubscribe from the TB-SUPPORT list, click the following link:
> https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
>
########################################################################
To unsubscribe from the TB-SUPPORT list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/webadmin?SUBED1=TB-SUPPORT&A=1
|