 |
 |
Thanks Wen On Wed, Sep 30, 2009 at 12:52 PM, Fabrizio Furano <[log in to unmask]> wrote: > Hi Wen, > > OK, I made a new source tarball: xrootd-20090930.1117 . It's available > here: > > https://savannah.cern.ch/projects/xrootd > > Just go to the "download" section. > > Please let me know if it works for you. > > Fabrizio > > > > wen guan wrote: >> >> Hi Fabrizio, >> >> Usually I got it for the xrootd binary tar file or the src >> file(compile it). If you can give me the src file, I can compile it. >> I checked my job. almost all of them crashed at the end. It seems they >> cannot read the last some hundreds events. I think maybe it's caused >> by a incompletely copied file. >> >> Regards >> Wen >> >> On Wed, Sep 30, 2009 at 11:49 AM, Fabrizio Furano <[log in to unmask]> wrote: >>> >>> It depends. Usually where do you get it from? >>> >>> f >>> >>> >>> wen guan wrote: >>>> >>>> Hi Fabrizio, >>>> >>>> Good! Thanks! >>>> Where can I get the new xrdcp for both 32 bit and 64 bit? >>>> >>>> Regards >>>> Wen >>>> >>>> On Wed, Sep 30, 2009 at 11:17 AM, Fabrizio Furano <[log in to unmask]> >>>> wrote: >>>>> >>>>> OK, I found it and fixed. Still I doubt that it worked in the older >>>>> version, >>>>> but anyway... >>>>> >>>>> The copied file was OK. It was just the final check on the number of >>>>> copied >>>>> bytes which was broken for files bigger than 4G. That is what code 13 >>>>> is >>>>> supposed to mean. >>>>> >>>>> Thank you for the precise report. >>>>> Fabrizio >>>>> >>>>> >>>>> >>>>> ---------- before >>>>> furano@pcitdm07 9:47:53 ~/Park/xrootd>./bin/xrdcp -S 15 >>>>> >>>>> >>>>> root://atlas-bkp2.cs.wisc.edu:1094//atlas/xrootd/atlasmcdisk/mc08/AOD/mc08.106051.PythiaZmumu_1Lepton.merge.AOD.e347_s462_s520_d153_r643_t53_tid078914/AOD.078914._000115.pool.root.1 >>>>> AOD.root >>>>> Set ParStreamsPerPhyConn to 15 >>>>> [xrootd] Total 4996.48 MB |====================| 100.00 % [8.3 >>>>> MB/s] >>>>> furano@pcitdm07 9:58:47 ~/Park/xrootd>echo $? >>>>> 13 >>>>> >>>>> >>>>> ---------- now >>>>> furano@pcitdm07 10:01:09 ~/Park/xrootd>./bin/xrdcp -S 15 >>>>> >>>>> >>>>> root://atlas-bkp2.cs.wisc.edu:1094//atlas/xrootd/atlasmcdisk/mc08/AOD/mc08.106051.PythiaZmumu_1Lepton.merge.AOD.e347_s462_s520_d153_r643_t53_tid078914/AOD.078914._000115.pool.root.1 >>>>> AOD2.root >>>>> Set ParStreamsPerPhyConn to 15 >>>>> [xrootd] Total 4996.48 MB |====================| 100.00 % [8.4 >>>>> MB/s] >>>>> furano@pcitdm07 10:11:48 ~/Park/xrootd>echo $? >>>>> 0 >>>>> >>>>> >>>>> >>>>> >>>>> Fabrizio Furano wrote: >>>>>> >>>>>> Hi Wen, >>>>>> >>>>>> actually if the result code is 13, then xrdcp thinks that it did not >>>>>> copy >>>>>> the right number of bytes. That looks strange to me, does it always >>>>>> happen >>>>>> with big files? Also, could you please post the lenghts of both the >>>>>> source >>>>>> and the destination files? >>>>>> >>>>>> Fabrizio >>>>>> >>>>>> wen guan wrote: >>>>>>> >>>>>>> Hi, >>>>>>> >>>>>>> When using this xrdcp to copy big files(about 5G), the return value >>>>>>> is not 0, which caused some program thought the copy failed and >>>>>>> crashed the job. Is it a bug? >>>>>>> xrootd-20080828 doesn't have this problem. when using >>>>>>> xrootd-20080828, >>>>>>> the return value is 0. >>>>>>> >>>>>>> >>>>>>> >>>>>>> [c105] /home/wguan > xrdcp >>>>>>> >>>>>>> >>>>>>> >>>>>>> root://atlas-bkp2.cs.wisc.edu:1094//atlas/xrootd/atlasmcdisk/mc08/AOD/mc08.106051.PythiaZmumu_1Lepton.merge.AOD.e347_s462_s520_d153_r643_t53_tid078914/AOD.078914._000115.pool.root.1 >>>>>>> AOD.root >>>>>>> [xrootd] Total 4996.48 MB |====================| 100.00 % [94.4 >>>>>>> MB/s] >>>>>>> [c105] /home/wguan > echo $? >>>>>>> 13 >>>>>>> >>>>>>> >>>>>>> Regards >>>>>>> Wen >