 |
 |
Hi Pete, Peter Elmer wrote: > Hi All, > > As we noted last week, some people are away at the supercomputing conference > this week so we will not have a live meeting. I do have some questions for > various people that hopefully can be answered by email... > > Andy: > > o Did you fix some dates to come to CERN? > > Fabrizio: > > o Remi posted something in the BaBar HN: > > http://babar-hn.slac.stanford.edu:5090/HyperNews/get/KanSOS/491.html > > about the remaining issues for using the xrootd/olbd production output > buffer. The outstanding issues appear to be: > > o setting the unix permissions correctly > o checksum doesn't work (might be server side config issue, see > my response in HN) > o file deletion doesn't seem to work > > o Could you summarize the other things (TXNetFile cleanup, refresh bug, > etc.) that are outstanding? > > Gerri/Fons: > > o Gerri mentioned something to me about some backward compability issue/bug > with the original rootd. Was that understood? > I think this has now been fixed by Gerri. > o Has anything else come up in (Alice) testing of xrootd? > We still plan to use it in the Computing Data Challenge where we will use it to write 100's TB of data. Also we are now using is to read data files on gLite storage elements that we use in the gLite-PROOF demo at SC'04. > o We realized that we really should try to get asynchronous mode working > before the next production ROOT release (in particular the one that will > be used with POOL) since it is required to start using unsolicited > responses. The problem is that if there are lots of clients out there > which are not running in asynchronous mode, it creates a significant > backward compatibility problem since they block when such a response > comes in. Fabrizio found that it works in the posix-like client, but > there is some issue with TThread/TSocket that causes problems for > TXNetFile. > We have made some corrections in the TThread class recently (including a port to Win32, where pthreads don't exists). So I would really like to urge Fabrizio and Gerri to give this another very intense try. Cheers, Fons. -- Org: CERN, European Laboratory for Particle Physics. Mail: 1211 Geneve 23, Switzerland E-Mail: [log in to unmask] Phone: +41 22 7679248 WWW: http://www.rademakers.org/fons/ Fax: +41 22 7679480