 |
 |
Hi Alvise, If you turn on debugging you should see this. Locks only make a difference when writing. Andy ----- Original Message ----- From: "Alvise Dorigo" <[log in to unmask]> To: "Andrew Hanushevsky" <[log in to unmask]> Cc: <[log in to unmask]> Sent: Wednesday, September 15, 2004 6:00 AM Subject: Re: [Fwd: errore xrootd] > Thanks Andy, > I do not see this for reading but only for writing (?). I've to investigate. > > BTW: I'll run soon olbd under gdb waiting for its crash and I'll send > you the stack trace. > > Alvise > > Andrew Hanushevsky wrote: > > Hi Alvise, > > > > This appears to be your problem :-) When you killed xrootd and then > > restarted it your client reconnected and reopened the file in write mode. > > Then your client (perhaps a different one or the same one) tried to open > > it again. So, xrootd did the right thing in this case. > > > > Andy > > > > On Tue, 14 Sep 2004, Alvise Dorigo wrote: > > > > > >> > >>Hi Andy; > >>I'm testing xrootd in writing mode (opening a file in "recreate" mode > >>which perfectly works). I killed xrootd during writing, then I resumed > >>it and when the client tried to reopen it in write mode I got this error > >>message: > >> > >>040914 17:28:48 9148 alvise.0:[log in to unmask] ofs_close: lnks=2 > >>pi=5 fn=/tmp/file > >>040914 17:28:48 9148 XrootdXeq: Output file /tmp/file is already opened > >>by 1 writer; open denied. > >>040914 17:28:48 9148 alvise.0:[log in to unmask] XrootdResponse: > >>0001 sending err 3003: Output file /tmp/file is already opened by 1 > >>writer; open denied. > >> > >>Is it supposed to act like this or is it an error ? > >> > >>thanks > >> > >> Alvise > >> > >> > >> > >> >