fcntl F_SETLKW64

[Romeril, Alan]

Okay, I`ll give it a go when the machine next does this, it might not be
for a while as it won`t be busy enough over the Christmas break.
And I doubt there will be enough time to catch one of the rouges in the
act as several smbds are doing this and most of these die by the time I
can attach a truss...


David Collier-Brown wrote:
> 
> Jeremy Allison wrote:
> >
> > On Fri, Dec 21, 2001 at 11:54:54AM +0000, Romeril, Alan wrote:
> >
> > > Also the fcntl F_SETLKW64 bug is still there I have just had to pkill
> > > the smbds AGAIN as the load ave hit 50.
> > > machine:bash-2.04#truss -p 13162
> > > fcntl(7, F_SETLKW64, 0xFFBEF7B0)                = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF7B0)                = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF7B0)                = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF820)                = 0
> > > kill(29956, SIG#0)                              = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF7B0)                = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF7B0)                = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF7B0)                = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF820)                = 0
> > > fcntl(7, F_SETLKW64, 0xFFBEF7B0)                = 0
> > >
> > > again yes Solaris 8, on an Ultra
> >
> > Can you get a backtrace when it's in that state so I can see
> > what smbd is doing ?
> 
> To get all the interesting data, try
>         pstack <pid>            -- get a stack traceback
> and then,
>         pfiles  <pid>           -- find out what fiels/sockets
>         pmap -x <pid>           -- memory map
> and
>         ptree -a <pid>          -- process tree
> 
> --dave
> 
> --
> David Collier-Brown,           | Always do right. This will gratify
> Americas Customer Engineering, | some people and astonish the rest.
> SunPS Integration Services.    |                      -- Mark Twain
> (905) 415-2849                 | davecb at canada.sun.com