ATLAS SC4 Site log for PHOENIX cluster

23. 6. 2006. Status after first week of tier-0 exercise

After the power cut at CERN, Tier-0 -> tier-1 transfers are now back up and pumping out 500 MB/s.

FZK has had some inital problems: "FZK was in scheduled downtime (until Tuesday evening), srm endpoint was not published in the BDII and FTS server could not find a correct channel." The downtime was extended until Wednesday evening because of gLite-3.0 upgrade. Current status unknown, but according to http://gridview.cern.ch/GRIDVIEW/, transfers have resumed, peaking at 130 MB/s.

Prague has been running initial dataset transfers, peaking at 50 MB/s (see http://golias100.farm.particle.cz/ipac/fzk_1/), but seing lots of unstability.

I have carried out initial transfer tests FZK -> CSCS with SRM and FTS. I have also submitted a request for CSCS to be added to the list of sites for DQ2 transfers (TiersOfAtlas). A technical log is kept at http://cern.ch/fjob/atlas/SC4-ATLAS-tests-CSCS-FZK.txt

In short, we are waiting for the situation CERN -> FZK (-> Prague) to stabilize, before we run throughput tests FZK -> CSCS.

-- FrederikOrellana - 23 Jun 2006